<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Ugo Tomasello on Medium]]></title>
        <description><![CDATA[Stories by Ugo Tomasello on Medium]]></description>
        <link>https://medium.com/@ugotomasello?source=rss-6bd5914fe82e------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*XQO8Sj09P5tDvazrUpWykw.png</url>
            <title>Stories by Ugo Tomasello on Medium</title>
            <link>https://medium.com/@ugotomasello?source=rss-6bd5914fe82e------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Sat, 21 Nov 2026 06:29:11 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@ugotomasello/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Your Brain in Milliseconds: The EEG Signals That Flag Seizures, Encephalopathy, and Coma]]></title>
            <link>https://medium.com/@ugotomasello/your-brain-in-milliseconds-the-eeg-signals-that-flag-seizures-encephalopathy-and-coma-a0e9f03369ff?source=rss-6bd5914fe82e------2</link>
            <guid isPermaLink="false">https://medium.com/p/a0e9f03369ff</guid>
            <category><![CDATA[brain]]></category>
            <category><![CDATA[mental-health]]></category>
            <category><![CDATA[neurology]]></category>
            <category><![CDATA[healthcare]]></category>
            <category><![CDATA[digital-biomarkers]]></category>
            <dc:creator><![CDATA[Ugo Tomasello]]></dc:creator>
            <pubDate>Sat, 03 Jan 2026 16:36:20 GMT</pubDate>
            <atom:updated>2026-01-03T16:42:41.742Z</atom:updated>
            <content:encoded><![CDATA[<p>EEG Biomarkers in Neurology: A Clinical-First Marker Library (PDR, spikes, LPDs, burst-suppression, EDB)</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*78rVq8atPwocA9fpkyh0Lw.png" /></figure><p>· <a href="#fee4">0. How to read this expanded library</a><br> ∘ <a href="#3066">Quick legend (the kinds of “markers” you’ll see):</a><br> ∘ <a href="#74e2">How neurology names work (PDR, LPDs, LRDA…):</a><br>· <a href="#2428">1. Background &amp; state markers (your EEG “vitals”)</a><br> ∘ <a href="#cde5">Posterior Dominant Rhythm (PDR / posterior alpha)</a><br> ∘ <a href="#1d29">Generalized slowing (theta/delta excess)</a><br> ∘ <a href="#08fe">Focal slowing</a><br>· <a href="#c139">2. Epileptiform (interictal) markers</a><br> ∘ <a href="#16d6">Spike (20–70 ms) / Sharp wave (70–200 ms)</a><br> ∘ <a href="#1c01">Spike-and-wave / Polyspike-and-wave</a><br>· <a href="#434c">3. Ictal markers + operational definitions (ICU + EMU)</a><br> ∘ <a href="#5095">Electrographic seizure (ACNS)</a><br> ∘ <a href="#1bbd">Electrographic status epilepticus (ESE; ACNS framework)</a><br>· <a href="#6f9b">4. Rhythmic/periodic ICU patterns (the ictal–interictal continuum)</a><br> ∘ <a href="#467d">LPDs / GPDs and LRDA / GRDA</a><br>· <a href="#3bb3">5. Encephalopathy signatures</a><br> ∘ <a href="#c02b">Triphasic waves</a><br>· <a href="#ea93">6. Coma &amp; neuroprognostication anchors (post–cardiac arrest)</a><br> ∘ <a href="#3a96">Suppressed background / Burst-suppression</a><br> ∘ <a href="#c16c">Alpha coma</a><br>· <a href="#2f9f">7. Sleep-stage markers relevant to neurology</a><br> ∘ <a href="#b83b">Spindles, K-complexes, slow waves</a><br>· <a href="#429f">8. Pediatric high-yield pattern</a><br> ∘ <a href="#7f4f">Hypsarrhythmia (infantile spasms / West syndrome)</a><br>· <a href="#e8ee">9. Autoimmune encephalitis signature</a><br> ∘ <a href="#6723">Extreme delta brush (anti-NMDAR encephalitis)</a><br>· <a href="#fd68">10. “Don’t-overtreat” benign variants (avoid false epilepsy labels)</a><br>· <a href="#f96f">11. Closing chapter: A reference you can actually use</a><br> ∘ <a href="#abf4">What this library is really saying</a><br> ∘ <a href="#d712">Why the significance goes beyond EEG readers, (If you zoom out, this library supports two big goals):</a><br>· <a href="#80e5">🧩 About me</a></p><h3>0. How to read this expanded library</h3><p>Following from the first episode about the EEG biomarkers in psychiatry, it’s time now to complement it with the neurological angle to study the brain functions. Neurology is the medical specialty focused on the <strong>nervous system, </strong>brain, spinal cord, peripheral nerves, neuromuscular junction, and muscles when something goes wrong in their <strong>structure or function</strong>. Neurologists diagnose and treat conditions like:</p><ul><li><strong>Stroke</strong> and transient ischemic attacks</li><li><strong>Epilepsy</strong> and seizures</li><li><strong>Multiple sclerosis</strong> and other inflammatory diseases</li><li><strong>Parkinson’s</strong> and movement disorders</li><li><strong>Migraine</strong> and headache syndromes</li><li><strong>Neuropathies</strong> (e.g., diabetic neuropathy), myasthenia, ALS, dementia</li></ul><p>In this context, neurology EEG is a bit different from psychiatry EEG. In psychiatry, we often use EEG as a <strong>probabilistic probe of circuits</strong> (attention, reward, salience). In neurology, EEG is frequently a <strong>clinical decision tool</strong>: it helps answer time-critical questions like “Is this a seizure?”, “How diffuse is brain dysfunction?”, or “Is the brain reacting?<em>”</em> (<a href="https://www.ncbi.nlm.nih.gov/books/NBK390343">NCBI</a>)</p><p>That’s why this library is “clinical-first.” Many entries are not subtle research features but <strong>named patterns</strong> with <strong>standardized reporting language</strong> (especially in the ICU, where the ACNS terminology exists to keep teams consistent). (<a href="https://www.acns.org/UserFiles/file/ACNSStandardizedCriticalCareEEGTerminology_rev2021.pdf">acns.org</a>)</p><h4><strong>Quick legend (the kinds of “markers” you’ll see):</strong></h4><ul><li><strong>Background/state markers:</strong> “vitals” of the EEG (organization, dominant rhythm, reactivity). (<a href="https://www.ncbi.nlm.nih.gov/books/NBK390343">NCBI</a>)</li><li><strong>Epileptiform transients:</strong> spikes/sharps that suggest seizure propensity (context matters). (<a href="https://www.ncbi.nlm.nih.gov/books/NBK557655">NCBI</a>)</li><li><strong>Rhythmic/periodic ICU patterns:</strong> patterns that can sit on the <strong>ictal–interictal continuum</strong>. (<a href="https://www.acns.org/UserFiles/file/ACNSStandardizedCriticalCareEEGTerminology_rev2021.pdf">acns.org</a>)</li><li><strong>Sleep graphoelements:</strong> normal sleep landmarks that can look “pathological” if you don’t recognize them. (<a href="https://www.ncbi.nlm.nih.gov/books/NBK537023">NCBI</a>)</li></ul><h4><strong>How neurology names work (PDR, LPDs, LRDA):</strong></h4><ul><li><strong>PDR</strong> = posterior dominant rhythm (your baseline “posterior alpha” in relaxed wakefulness). (<a href="https://www.ncbi.nlm.nih.gov/books/NBK390343/">NCBI</a>)</li><li><strong>LPDs/GPDs</strong> = lateralized/generalized periodic discharges (ACNS “main terms”).</li><li><strong>LRDA/GRDA</strong> = lateralized/generalized rhythmic delta activity. (<a href="https://pmc.ncbi.nlm.nih.gov/articles/PMC10508310">PMC</a>)</li><li>In ICU language, you often report frequency<strong> (Hz), evolution, reactivity, plus-modifiers, and burden</strong>.</li></ul><h3>1. Background &amp; state markers (your EEG “vitals”)</h3><h4>Posterior Dominant Rhythm (PDR / posterior alpha)</h4><p><strong>What it is:</strong> The classic relaxed, eyes-closed wake rhythm, posterior, fairly symmetric, and <strong>reactive</strong> (it attenuates with eye opening/alerting).</p><p><strong>How clinicians “score” it (informally but consistently):</strong></p><ul><li><strong>Frequency</strong> (is it age-appropriate?)</li><li><strong>Symmetry</strong> (one side weaker/slower?)</li><li><strong>Reactivity</strong> (does it block with eye opening or stimulation?)</li></ul><p><strong>Why it matters:</strong> If the PDR is slowed, poorly organized, or non-reactive, it often points toward <strong>diffuse cerebral dysfunction</strong>, especially when paired with generalized slowing. (<a href="https://www.ncbi.nlm.nih.gov/books/NBK564371/">NCBI</a>)</p><p><strong>Common pitfall:</strong> Mistaking “alpha coma” for a normal posterior alpha rhythm. Alpha coma is typically <strong>non-reactive</strong> and occurs in coma states, reactivity is the key discriminator. (<a href="https://www.ncbi.nlm.nih.gov/books/NBK539805">NCBI</a>)</p><h4>Generalized slowing (theta/delta excess)</h4><p><strong>What it is:</strong> Diffuse slowing of the background (more theta/delta, less organized posterior rhythm).</p><p><strong>Clinical meaning:</strong> Often reflects <strong>global brain dysfunction</strong> (toxic-metabolic, medication/sedation effects, anoxic injury), and the pattern’s severity generally tracks encephalopathy depth.</p><h4>Focal slowing</h4><p><strong>What it is:</strong> Persistent slow activity localized to a region.</p><p><strong>Clinical meaning:</strong> Raises suspicion for a <strong>focal structural/functional disturbance</strong> (stroke, mass, focal encephalitis), especially if persistent and spatially consistent.</p><h3>2. Epileptiform (interictal) markers</h3><h4>Spike (20–70 ms) / Sharp wave (70–200 ms)</h4><p><strong>What it is:</strong> Brief epileptiform transients — often with a convincing <strong>field</strong> and sometimes an after-going slow wave. (Exact millisecond cutoffs matter less than morphology + field + context.)</p><p><strong>How it’s described in practice:</strong></p><ul><li><strong>Where</strong> (focal vs generalized; which region)</li><li><strong>How often</strong> (rare/occasional/frequent)</li><li><strong>State dependence</strong> (wake vs drowsy/sleep)</li><li><strong>Whether it’s truly epileptiform</strong> (field, sharpness, disruption of background) (<a href="https://www.ncbi.nlm.nih.gov/books/NBK558912">NCBI</a>)</li></ul><p><strong>Clinical interpretation (careful):</strong></p><ul><li>Interictal epileptiform discharges can support epilepsy risk, but they are not an automatic “seizure diagnosis” without clinical correlation.</li><li>Some normal variants mimic reminding shapes, see Section 10. <a href="https://www.ncbi.nlm.nih.gov/books/NBK555899/">NCBI</a></li></ul><h4>Spike-and-wave / Polyspike-and-wave</h4><p><strong>What it is:</strong> Classic generalized epilepsy signatures in the right syndrome/context.</p><p><strong>Where it shows up:</strong> ILAE descriptions link generalized spike-wave and polyspike-wave patterns to idiopathic/genetic generalized epilepsy syndromes and seizure types (e.g., absence, myoclonic). (<a href="https://www.ilae.org/files/dmfile/idiopathic-generalized2.pdf">ilae.org</a>)</p><p><strong>Pitfall:</strong> Overcalling generalized patterns without confirming <strong>true generalization</strong> (synchronous, bilateral onset) and without considering benign look-alikes or medication effects. (<a href="https://www.ncbi.nlm.nih.gov/books/NBK555899/">NCBI</a>)</p><h3>3. Ictal markers + operational definitions (ICU + EMU)</h3><p>Here’s where neurology EEG becomes “rule-based,” because decisions (treat, escalate, monitor) depend on shared definitions.</p><h4>Electrographic seizure (ACNS)</h4><p><strong>Definition (simplified):</strong> Either</p><ul><li>Epileptiform discharges averaging <strong>&gt;2.5 Hz for ≥10 seconds</strong> (≥25 discharges/10 s), <strong>or</strong></li><li>A pattern with <strong>definite evolution</strong> lasting <strong>≥10 seconds</strong>. (<a href="https://abret.org/index.php/download_file/954">ABRET</a>)</li></ul><p><strong>Why it matters:</strong> This prevents ambiguous language (e.g., calling a clear seizure “periodic discharges”), especially when different clinicians hand off care.</p><h4>Electrographic status epilepticus (ESE; ACNS framework)</h4><p><strong>Core idea:</strong> Seizure activity that is <strong>prolonged</strong> or <strong>dominates a large fraction of time, </strong>definitions are used to standardize ICU decision-making.</p><p><strong>Practical takeaway:</strong> If you’re building protocols or products, these definitions are your “API contract”: they align clinicians, EEG readers, and downstream actions.</p><h3>4. Rhythmic/periodic ICU patterns (the ictal–interictal continuum)</h3><p>ICU EEG has many patterns that are not cleanly “seizure” vs “not seizure.” ACNS terminology keeps them readable and comparable.</p><h4>LPDs / GPDs and LRDA / GRDA</h4><p><strong>What they are:</strong> Standard ACNS “main terms” for periodic discharges and rhythmic delta activity.</p><p><strong>How to report them (minimum useful set):</strong></p><ul><li><strong>Frequency</strong> (Hz categories), <strong>evolution</strong>, and <strong>plus-modifiers</strong> (“more ictal-appearing”).</li><li><strong>Reactivity</strong> and <strong>burden</strong> (how much of the record is occupied).</li></ul><p><strong>Why clinicians care:</strong> Seizure risk is not uniform:</p><ul><li>Evidence syntheses and large datasets show that patterns like <strong>LPDs</strong> and <strong>LRDA</strong> are often associated with seizures, and risk tends to rise with faster frequency and “plus” features. (<a href="https://link.springer.com/article/10.1007/s10309-022-00526-4">Springer Nature</a>)</li></ul><p><strong>Pitfall:</strong> Treating every IIC pattern as “status.” Many are <strong>risk markers</strong> that require clinical correlation, not automatic escalation.</p><h3>5. Encephalopathy signatures</h3><h4>Triphasic waves</h4><p><strong>What they look like:</strong> High-amplitude, sharply contoured waves with <strong>three phases</strong> — classically taught as “hepatic,” but now recognized as <strong>non-specific</strong>. (<a href="https://www.ncbi.nlm.nih.gov/books/NBK557679">NCBI</a>)</p><p><strong>What they mean (modern view):</strong> Triphasic morphology can appear across metabolic/toxic encephalopathies and even structural etiologies; treat it as a clue to <strong>diffuse dysfunction</strong>, not a single-cause stamp.</p><p><strong>Pitfall:</strong> Over-interpreting triphasic waves as one diagnosis — or missing that some periodic patterns with triphasic morphology can sit near the IIC. ACNS language helps you separate morphology from seizure criteria.</p><h3>6. Coma &amp; neuroprognostication anchors (post–cardiac arrest)</h3><p>This is where EEG carries heavy ethical weight, so guidelines emphasize timing and confounders.</p><h4>Suppressed background / Burst-suppression</h4><p><strong>Why it matters:</strong> The Neurocritical Care Society guideline suggests that <strong>suppressed or burst-suppression background (± periodic discharges)</strong> on EEG performed <strong>≥72 hours</strong> from ROSC (or ≥72 hours from rewarming after hypothermia), <strong>without sedation or other confounders</strong>, is a <strong>predictor of poor functional outcome</strong> with low false-positive rates in the evidence base. (<a href="https://www.neurocriticalcare.org/Portals/0/Guidelines/Neuroprognostication_in_comatose_survivors_of_cardiac_arrest_Neurocritical_Care_March_2023.pdf">Neurocritical Care Society</a>)</p><p><strong>Key nuance:</strong> These patterns can appear early after arrest and later evolve into less malignant patterns; sedation/hypothermia can also induce similar appearances — hence the emphasis on timing.</p><h4>Alpha coma</h4><p><strong>What it is:</strong> EEG dominated by alpha-range activity in coma, typically <strong>non-reactive</strong> — not to be confused with normal posterior alpha. (<a href="https://www.ncbi.nlm.nih.gov/books/NBK539805">NCBI</a>)</p><h3>7. Sleep-stage markers relevant to neurology</h3><p>Sleep is where normal physiology can masquerade as pathology — and where some epilepsies become louder.</p><h4>Spindles, K-complexes, slow waves</h4><p><strong>What they are:</strong> Core staging landmarks (N2 spindles/K-complexes; N3 slow waves). (<a href="https://www.ncbi.nlm.nih.gov/books/NBK537023">NCBI</a>)</p><p><strong>Why they matter clinically:</strong></p><ul><li>Many epileptiform discharges are <strong>sleep-activated</strong>.</li><li>Recognizing normal sleep elements prevents false “seizure” labeling (especially with benign variants that cluster in drowsiness).</li></ul><h3>8. Pediatric high-yield pattern</h3><h4>Hypsarrhythmia (infantile spasms / West syndrome)</h4><p><strong>What it looks like:</strong> Chaotic, high-voltage slow background with multifocal spikes — often described as “chaos,” and it can be incomplete or intermittent early. (<a href="https://pmc.ncbi.nlm.nih.gov/articles/PMC9985829">PMC</a>)</p><p><strong>Clinical framing:</strong> Infantile spasms are treated as an emergency; EEG can be decisive, but repeating/extended recordings may be needed if early tracings are equivocal. (<a href="https://practicalneurology.com/diseases-diagnoses/epilepsy-seizures/epilepsy-essentials-infantile-spasms-the-biggest-little-seizures/31995">practicalneurology.com</a>)</p><h3>9. Autoimmune encephalitis signature</h3><h4>Extreme delta brush (anti-NMDAR encephalitis)</h4><p><strong>What it is:</strong> Rhythmic delta with superimposed fast activity (“brush”), a striking pattern seen in a subset of adults with anti-NMDAR encephalitis. (<a href="https://pmc.ncbi.nlm.nih.gov/articles/PMC3525298">PMC</a>)</p><p><strong>Why it’s useful:</strong> In the original adult series, its presence was associated with <strong>more prolonged illness</strong>; it’s best treated as a supportive pattern in the right clinical context, not a standalone diagnostic test. (<a href="https://pubmed.ncbi.nlm.nih.gov/22933737">PubMed</a>)</p><h3>10. “Don’t-overtreat” benign variants (avoid false epilepsy labels)</h3><p>Some patterns look sharp, rhythmic, or “spiky” and still are normal variants.</p><p><strong>High-yield examples:</strong> wicket waves, benign epileptiform transients of sleep (BETS/SSS), 14&amp;6 positive spikes, SREDA, phantom 6-Hz spike-wave, rhythmic temporal theta bursts of drowsiness.</p><p><strong>How to keep your footing:</strong></p><ul><li>Check <strong>state</strong> (drowsiness/light sleep is a hotspot for benign variants).</li><li>Look for lack of <strong>evolution</strong> and absence of convincing epileptiform field/after-going slow wave.</li><li>When uncertain, use conservative language and correlate clinically.</li></ul><h3>11. Closing chapter: A reference you can actually use</h3><p>If psychiatry EEG is often about dimensions, neurology EEG is often about decisions. The library you’ve just read is less a catalog of interesting waveforms than a practical translation layer between what the brain is doing and what a clinical team needs to decide next: Is the brain globally slowed or locally injured? Is this seizure activity, seizure risk, or neither? Is the EEG reactive? Is the ICU pattern drifting toward harm? so the next clinician (or your future self) can interpret it safely.</p><p>The deepest lesson is simple: EEG patterns gain meaning through <strong>context + standardization</strong>. Context is the patient (sedation, hypoxia, fever, liver failure). Standardization is the language (ACNS terms, seizure definitions, neuroprognostication timing). When you combine the two, EEG becomes less of an art object and more of a reliable clinical instrument — fast, repeatable, and often life-saving.</p><h4>What this library is really saying</h4><ul><li><strong>Start with the EEG “vitals.”</strong><br>Before you hunt for anything fancy, check the basics: Is the background rhythm organized? Is there a normal posterior dominant rhythm (PDR) when the person is relaxed with eyes closed? Does it fade when they open their eyes or when you stimulate them (reactivity)? These simple checks often tell you whether the brain looks broadly “awake and organized” or more globally affected.</li><li><strong>Don’t mix up “warning signs” with “events.”</strong><br>Spikes and sharp waves can mean the brain has an increased tendency toward seizures, but they are not the same thing as an actual seizure happening right now. Think of them like “risk signals,” not proof of an ongoing event, unless the pattern meets seizure criteria. (<a href="https://www.learningeeg.com/epileptiform-activity">Learning EEG</a>)</li><li><strong>Use a shared ICU vocabulary so everyone means the same thing.</strong><br>In critical care, EEG readers use standardized labels (like <strong>LPDs/GPDs</strong> and <strong>LRDA/GRDA</strong>) plus simple descriptors (how fast, whether it changes over time, and how much of the recording it occupies). This isn’t jargon for its own sake, it prevents miscommunication when teams hand off patients and when different hospitals compare findings.</li><li><strong>Anchor big decisions to clear rules.</strong><br>When the question is “Is this a seizure?”, there are agreed operational definitions (for example: discharges faster than <strong>2.5 per second</strong> lasting <strong>10 seconds</strong>, or a pattern that clearly evolves). Using these rules keeps treatment decisions consistent and makes ICU EEG much more reliable across readers and sites</li></ul><h4><strong>Why the significance goes beyond EEG readers, </strong>(If you zoom out, this library supports two big goals):</h4><ul><li><strong>Safer clinical decisions (less over- and under-treatment):</strong></li><li>Recognizing benign variants prevents false epilepsy labels and unnecessary antiseizure medications.</li><li>Treating IIC patterns as <em>risk signals</em> (not automatic “status”) encourages appropriately staged responses, repeat exams, imaging correlation, and targeted treatment when warranted.</li><li><strong>Better systems of care (monitoring that matches risk):</strong></li><li>Continuous EEG is recommended for detecting nonconvulsive seizures/status and for assessing treatment response — often until seizures are controlled for at least 24 hours.</li><li>Practical tools like <strong>2HELPS2B</strong> help decide how long to monitor and how worried to be after early EEG findings, turning “intuition” into a repeatable workflow.</li></ul><h3>🧩 About me</h3><p>With a background spanning Cognitive Neurosciences, <strong>R&amp;D in Molecular Neurobiology</strong>, Clinical Development Management, I am earning my stripes in Venture Capital, because I love to explore new perspectives to evaluate Tech innovation and <strong>investments in Life Sciences</strong>. My commitment to transforming scientific insights into tangible health-tech solutions fuels my passion for this evolving AI-healthcare intersection especially when related to the brain and mental health.</p><p>If you want to discuss these exciting trends further — please feel free to reach out on <a href="https://www.linkedin.com/in/ugo-tomasello/">LinkedIn</a> or via <a href="http://ugo.tomasello@gmail.com/">email</a>.</p><p>👉 <strong>You can find me on </strong><a href="https://www.linkedin.com/in/ugo-tomasello/"><strong>LinkedIn</strong></a><strong> and keep up to date with my latest insights on </strong><a href="https://medium.com/@ugotomasello"><strong>Medium</strong></a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=a0e9f03369ff" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Your Brain in Milliseconds: The EEG Signals That Hint at Anxiety, Depression, and Psychosis]]></title>
            <link>https://medium.com/@ugotomasello/your-brain-in-milliseconds-the-eeg-signals-that-hint-at-anxiety-depression-and-psychosis-a2bd720d8aef?source=rss-6bd5914fe82e------2</link>
            <guid isPermaLink="false">https://medium.com/p/a2bd720d8aef</guid>
            <category><![CDATA[digital-biomarkers]]></category>
            <category><![CDATA[healthcare]]></category>
            <category><![CDATA[electroencephalography]]></category>
            <category><![CDATA[psychiatry]]></category>
            <category><![CDATA[braintech]]></category>
            <dc:creator><![CDATA[Ugo Tomasello]]></dc:creator>
            <pubDate>Fri, 02 Jan 2026 21:51:03 GMT</pubDate>
            <atom:updated>2026-01-02T21:51:03.306Z</atom:updated>
            <content:encoded><![CDATA[<p><em>EEG Biomarkers in Psychiatry: A Quick-Check Marker Library (P50, MMN, P300, RewP)</em></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*_CUh4P8L0N3IYJqe-HtS0g.png" /></figure><p>· <a href="#dc29">0. How to read this expanded library</a><br>· <a href="#bb01">1. Sensory gating &amp; early encoding (input filtering)</a><br> ∘ <a href="#d2f0">P50 sensory gating (paired-click)</a><br> ∘ <a href="#68eb">N100 sensory encoding / gating</a><br>· <a href="#2371">2. Predictive processing (automatic deviance detection)</a><br> ∘ <a href="#ac46">Mismatch Negativity (MMN; ~100–250 ms)</a><br>· <a href="#5e82">3. Attention, working memory, and cognitive speed</a><br> ∘ <a href="#7742">P300 / P3b (oddball; ~300 ms)</a><br> ∘ <a href="#36c6">N2 (go/no-go, flanker; ~200–350 ms)</a><br> ∘ <a href="#91f4">CNV (S1–S2 expectancy)</a><br>· <a href="#d102">4. Performance monitoring &amp; learning from outcomes</a><br> ∘ <a href="#c2a0">ERN (error-related negativity; response-locked)</a><br> ∘ <a href="#38dc">Pe (error positivity)</a><br> ∘ <a href="#b486">FRN (feedback-related negativity; ~200–350 ms post-feedback)</a><br> ∘ <a href="#ef8a">Feedback-P3</a><br>· <a href="#2850">5. Reward and motivation (positive valence circuits)</a><br> ∘ <a href="#77c3">RewP / reward positivity (~250–350 ms; gain–loss contrast)</a><br>· <a href="#908a">6. Emotion reactivity and regulation (negative valence circuits)</a><br> ∘ <a href="#ca83">LPP (late positive potential; sustained)</a><br>· <a href="#6ecc">7. Language &amp; thought organization</a><br> ∘ <a href="#4cea">N400 (semantic expectancy; ~300–500 ms)</a><br>· <a href="#af4e">8. Social cognition</a><br> ∘ <a href="#a01d">N170 / VPP (face processing; ~140–200 ms)</a><br>· <a href="#40aa">9. Oscillatory “circuit probes” (E/I balance, coordination)</a><br> ∘ <a href="#dfa1">40-Hz ASSR (auditory steady-state; power + phase-locking)</a><br> ∘ <a href="#6882">Resting frontal alpha asymmetry (FAA)</a><br>· <a href="#b3f6">10. Network dynamics &amp; “whole-brain” summaries</a><br> ∘ <a href="#1916">EEG microstates (A–D temporal parameters)</a><br>· <a href="#48e0">11. Sleep EEG endophenotypes with psychiatric relevance</a><br> ∘ <a href="#e3f8">Sleep spindles (sigma; density/amplitude; NREM Stage 2)</a></p><h3>0. How to read this expanded library</h3><p>I compiled a list of meaningful psychiatry markers for non-experts seeking guidance in the EEG biomarkers. Keep in mind that these markers are best treated as <strong>probabilistic probes of circuit function</strong> instead of standalone points for diagnoses.</p><p>Being the brain a highly interconnected system with complex dynamics, the value of these biomarkers increases sharply when you <strong>lock down the paradigm (stimuli + timing), preprocessing (filters/ICA), scoring window, and brain topography</strong>, and when you interpret them as <em>dimensions</em> of function (gating, prediction error, control, salience, and reward).</p><p>Before we dive into the library, here a quick legend to explain common terms used in EEG field.</p><p><strong>What kind of EEG “marker” are we talking about?</strong></p><ul><li><strong>ERP (Event-Related Potential):</strong> a small wave in the EEG that appears <strong>right after something happens</strong> (a sound, an image, a button press). You average many repeats to see it clearly.</li><li><strong>Brain rhythms (frequency markers):</strong> instead of a “wave after an event,” you look at <strong>how strong certain brain rhythms are</strong>, like <em>alpha</em> or <em>gamma</em> (measured in <strong>Hz</strong>, meaning cycles per second).</li><li><strong>Whole-brain patterns:</strong> summaries of <strong>overall activity shapes</strong> across the scalp (useful for “brain state” style descriptions).</li></ul><p><strong>How marker names work (P50, N100, P300…)</strong></p><ul><li><strong>P</strong> means the wave goes <strong>up</strong> (positive). <strong>N</strong> means it goes <strong>down</strong> (negative).</li><li>The number is roughly <strong>when it happens</strong>, in <strong>milliseconds (ms)</strong> after the event.</li><li>Example: <strong>P300</strong> = a positive wave around <strong>300 ms</strong> after a stimulus.</li></ul><p><strong>Common task labels (what people do during the EEG)</strong></p><ul><li><strong>Paired-click:</strong> two quick sounds (<strong>S1</strong> then <strong>S2</strong>) to test “filtering” of repeated input.</li><li><strong>Oddball:</strong> many common sounds/images + a few rare ones to test noticing “something different.”</li><li><strong>Go/No-Go / Flanker:</strong> tasks that test <strong>self-control</strong> and <strong>conflict handling</strong>.</li><li><strong>Feedback / reward tasks:</strong> you see outcomes (win/loss) to study <strong>reward and learning signals</strong>.</li></ul><p><strong>How researchers measure a marker</strong></p><ul><li><strong>Amplitude:</strong> how <strong>big</strong> the EEG response is.</li><li><strong>Latency:</strong> how <strong>fast</strong> it happens (timing).</li><li><strong>Difference score:</strong> compare two conditions (e.g., <em>rare — common</em>).</li><li><strong>Ratio:</strong> compare two parts of a task (common in gating tasks).</li></ul><p><strong>Where it appears on the head (topography)</strong></p><ul><li>Terms like <strong>front</strong>, <strong>center</strong>, <strong>back</strong> mean <strong>which scalp areas</strong> show the strongest signal.</li><li>Always note <strong>reference montage</strong> (average, mastoids, etc.) because it can change appearance.</li></ul><p>Now that we have shared some common references, let’s dive into this neuropsychiatry marker library collection!</p><h3>1. Sensory gating &amp; early encoding (input filtering)</h3><h4>P50 sensory gating (paired-click)</h4><ul><li><strong>What it is:</strong> A short-latency ERP (~50 ms) reflecting early auditory filtering.</li><li><strong>How it’s elicited/scored:</strong> Two clicks <strong>500 ms apart</strong> (S1/conditioning, S2/test). “Gating” is typically summarized as <strong>T/C ratio</strong>, with <strong>&lt;~50%</strong> often treated as “normal” suppression. (<a href="https://pmc.ncbi.nlm.nih.gov/articles/PMC7426515/">pmc.ncbi.nlm.nih.gov</a>)</li><li><strong>Clinical pattern:</strong> <strong>Higher T/C ratios</strong> (weaker suppression) are consistently reported in schizophrenia-spectrum samples, including meta-analytic work; it’s also not fully diagnostically specific (reported in other conditions too). (<a href="https://pmc.ncbi.nlm.nih.gov/articles/PMC7426515/">pmc.ncbi.nlm.nih.gov</a>)</li><li><strong>Mechanistic hooks:</strong> Linked to <strong>GABA_B-mediated inhibition</strong> and genetic associations (e.g., CHRNA7-related work discussed in reviews), but <strong>can be medication-sensitive</strong> (e.g., atypical antipsychotic effects), which complicates “trait marker” claims. (<a href="https://pmc.ncbi.nlm.nih.gov/articles/PMC7426515/">pmc.ncbi.nlm.nih.gov</a>)</li><li><strong>Pitfalls:</strong> Reported <strong>test–retest reliability can be modest</strong> and varies by pipeline; gating ratios can move with state, medication, and recording quality. (<a href="https://pmc.ncbi.nlm.nih.gov/articles/PMC7426515/">pmc.ncbi.nlm.nih.gov</a>)</li></ul><h4>N100 sensory encoding / gating</h4><ul><li><strong>What it is:</strong> Early auditory encoding response around <strong>~100 ms</strong>, often scored in the same paired-click paradigm (as “N100 gating”). (<a href="https://pmc.ncbi.nlm.nih.gov/articles/PMC7426515/">pmc.ncbi.nlm.nih.gov</a>)</li><li><strong>Clinical pattern:</strong> N100 gating deficits have also been reported in schizophrenia, and N100 is frequently analyzed alongside P50/MMN as part of an “early auditory processing” (EAP) battery. (<a href="https://pmc.ncbi.nlm.nih.gov/articles/PMC7426515/">pmc.ncbi.nlm.nih.gov</a>)</li></ul><h3>2. Predictive processing (automatic deviance detection)</h3><h4>Mismatch Negativity (MMN; ~100–250 ms)</h4><ul><li><strong>What it is:</strong> A pre-attentive “prediction error” response to deviant stimuli within a regular sequence (auditory oddball).</li><li><strong>Why it’s favored:</strong> MMN is one of the most replicated ERP abnormalities in schizophrenia; beyond group differences, it often correlates with <strong>cognition and functional outcomes</strong> (daily functioning/independent living proxies). (<a href="https://pmc.ncbi.nlm.nih.gov/articles/PMC4775447/?utm_source=chatgpt.com">pmc.ncbi.nlm.nih.gov</a>)</li><li><strong>Stimulus-type nuance (important):</strong> Reviews highlight that <strong>duration/intensity</strong> deviants may show abnormalities earlier, while <strong>frequency</strong> deviance effects may become more prominent later in illness course — so “MMN” is not one thing unless you specify the deviant. (<a href="https://pmc.ncbi.nlm.nih.gov/articles/PMC7426515/">pmc.ncbi.nlm.nih.gov</a>)</li><li><strong>Actionability:</strong> Higher MMN (or less-impaired MMN) has been reported to predict <strong>better response to auditory perceptual training</strong> in some studies — useful for enrichment/stratification hypotheses in trials. (<a href="https://pmc.ncbi.nlm.nih.gov/articles/PMC7426515/">pmc.ncbi.nlm.nih.gov</a>)</li><li><strong>Pitfalls:</strong> Mixed findings for endophenotype claims in relatives/high-risk groups in some reviews — so it’s strong as a <strong>group-level</strong> marker, less settled as an individual-level risk screen. (<a href="https://pmc.ncbi.nlm.nih.gov/articles/PMC7426515/">pmc.ncbi.nlm.nih.gov</a>)</li></ul><h3>3. Attention, working memory, and cognitive speed</h3><h4>P300 / P3b (oddball; ~300 ms)</h4><ul><li><strong>What it is:</strong> A classic marker of <strong>attentional allocation/context updating</strong>.</li><li><strong>Interpretation knobs: Amplitude</strong> tends to track engagement/available cognitive resources. <strong>Latency</strong> is commonly treated as an index of processing speed. (<a href="https://pmc.ncbi.nlm.nih.gov/articles/PMC5901243/">pmc.ncbi.nlm.nih.gov</a>)</li><li><strong>What to report:</strong> oddball parameters (target probability, inter-stimulus interval), reference montage, and whether you’re quantifying <strong>P3b vs P3a</strong> (novelty/attention shift).</li></ul><h4>N2 (go/no-go, flanker; ~200–350 ms)</h4><ul><li><strong>What it is:</strong> A family of negativities linked to <strong>cognitive control</strong>, often prominent on <strong>no-go</strong> trials and in conflict tasks; literature emphasizes that multiple processes can contribute (control + mismatch/novelty depending on design). (<a href="https://pmc.ncbi.nlm.nih.gov/articles/PMC4033096/">pmc.ncbi.nlm.nih.gov</a>)</li><li><strong>Practical note:</strong> N2 is highly task-dependent; label it by paradigm (e.g., “no-go N2”) and avoid over-interpreting as a single construct. (<a href="https://pubmed.ncbi.nlm.nih.gov/17850238/?utm_source=chatgpt.com">PubMed</a>)</li></ul><h4>CNV (S1–S2 expectancy)</h4><ul><li><strong>What it is:</strong> A slow negative shift between warning cue (S1) and target (S2), commonly interpreted as <strong>anticipation/preparation</strong> and sustained attention.</li><li><strong>Why it matters:</strong> CNV provides a “continuous” measure of preparatory control (not just a peak), useful for ADHD/control paradigms and intervention studies where preparation is the target.</li></ul><h3>4. Performance monitoring &amp; learning from outcomes</h3><h4>ERN (error-related negativity; response-locked)</h4><ul><li><strong>What it is:</strong> A rapid error-monitoring signal (often fronto-central) associated with performance monitoring/control.</li><li><strong>Clinical pattern:</strong> Meta-analytic work shows <strong>anxiety is associated with enhanced ERN</strong> (small-to-medium overall), with <strong>worry/anxious apprehension</strong> showing stronger effects than other anxiety dimensions. (<a href="https://pmc.ncbi.nlm.nih.gov/articles/PMC3744033/">pmc.ncbi.nlm.nih.gov</a>)</li></ul><h4>Pe (error positivity)</h4><ul><li><strong>What it is:</strong> A later component often linked to <strong>error awareness/processing</strong>; typically analyzed with ERN to separate “rapid monitoring” from “conscious evaluation.” (Widely used in performance-monitoring batteries; interpret cautiously as it’s sensitive to task demands and awareness manipulations.)</li></ul><h4>FRN (feedback-related negativity; ~200–350 ms post-feedback)</h4><ul><li><strong>What it is:</strong> A feedback evaluation response often discussed in reinforcement-learning terms (ACC/dopamine teaching signals). (<a href="https://pmc.ncbi.nlm.nih.gov/articles/PMC4033096/">pmc.ncbi.nlm.nih.gov</a>)</li><li><strong>Key nuance:</strong> Evidence suggests FRN can behave like a <strong>“more-or-less than expected”</strong> signal rather than purely “worse-than-expected,” and it is modulated by how expectations are formed (explicit cues vs learned). (<a href="https://pmc.ncbi.nlm.nih.gov/articles/PMC4033096/">pmc.ncbi.nlm.nih.gov</a>)</li><li><strong>Pitfall:</strong> Avoid equating FRN = “negative valence” without modeling expectancy; consider computational RL regressors when possible. (<a href="https://pmc.ncbi.nlm.nih.gov/articles/PMC4033096/">pmc.ncbi.nlm.nih.gov</a>)</li></ul><h4>Feedback-P3</h4><ul><li><strong>What it is:</strong> A later evaluative/updating component following feedback, often complementing FRN (FRN = rapid evaluation; P3 = updating/context revision).</li></ul><h3>5. Reward and motivation (positive valence circuits)</h3><h4>RewP / reward positivity (~250–350 ms; gain–loss contrast)</h4><ul><li><strong>What it is:</strong> An ERP difference often interpreted as <strong>reward responsiveness</strong>.</li><li><strong>State of evidence:</strong> The literature supports RewP as a <strong>candidate</strong> depression risk/symptom marker, but with heterogeneity and reproducibility concerns. (<a href="https://pmc.ncbi.nlm.nih.gov/articles/PMC10863647/?utm_source=chatgpt.com">pmc.ncbi.nlm.nih.gov</a>)</li><li><strong>Selective reporting / power reality-check:</strong> A registered-report p-curve analysis found <strong>weak evidential value</strong> and <strong>low average power (~20–27%)</strong> across ERN/RewP studies — meaning effects may exist, but many studies are underpowered and results can be unstable. (<a href="https://pubmed.ncbi.nlm.nih.gov/31987869/">PubMed</a>)</li><li><strong>Practical implication:</strong> Pre-register scoring/topography, standardize task and contrasts, and treat RewP as one feature in a panel rather than a single KPI. (<a href="https://pubmed.ncbi.nlm.nih.gov/31987869/">PubMed</a>)</li></ul><h3>6. Emotion reactivity and regulation (negative valence circuits)</h3><h4>LPP (late positive potential; sustained)</h4><ul><li><strong>What it is:</strong> A slow positive waveform emerging around <strong>~200–300 ms</strong> for emotional stimuli (often parietal in adults; more occipital in children), reflecting <strong>facilitated attention to motivationally salient content</strong>. (<a href="https://pmc.ncbi.nlm.nih.gov/articles/PMC4449331/">pmc.ncbi.nlm.nih.gov</a>)</li><li><strong>Regulation sensitivity:</strong> When participants <strong>reappraise</strong> (reinterpret) emotional stimuli, LPP amplitude typically <strong>reduces</strong>, and the magnitude of modulation can track subjective arousal. (<a href="https://pmc.ncbi.nlm.nih.gov/articles/PMC4449331/">pmc.ncbi.nlm.nih.gov</a>)</li><li><strong>Depression angle:</strong> Studies show LPP can be <strong>blunted</strong> with higher depressive symptoms, consistent with reduced engagement with emotional material. (<a href="https://pmc.ncbi.nlm.nih.gov/articles/PMC11292834/">pmc.ncbi.nlm.nih.gov</a>)</li><li><strong>Pitfalls:</strong> Developmental topography differs; separate “reactivity” (early window) from “regulation” (post-instruction window) when using reappraisal designs. (<a href="https://pmc.ncbi.nlm.nih.gov/articles/PMC11292834/">pmc.ncbi.nlm.nih.gov</a>)</li></ul><h3>7. Language &amp; thought organization</h3><h4>N400 (semantic expectancy; ~300–500 ms)</h4><ul><li><strong>What it is:</strong> A robust marker of <strong>meaning processing</strong> elicited by words and many nonverbal meaning-laden stimuli; <strong>latency is relatively stable</strong>, while <strong>amplitude</strong> is highly sensitive to context/expectancy and semantic access. (<a href="https://pmc.ncbi.nlm.nih.gov/articles/PMC4052444/">pmc.ncbi.nlm.nih.gov</a>)</li><li><strong>Why it’s useful clinically:</strong> N400 paradigms can probe semantic prediction/integration — relevant to disorganized thought/language phenotypes — without relying on self-report. (<a href="https://pmc.ncbi.nlm.nih.gov/articles/PMC4052444/">pmc.ncbi.nlm.nih.gov</a>)</li></ul><h3>8. Social cognition</h3><h4>N170 / VPP (face processing; ~140–200 ms)</h4><ul><li><strong>What it is:</strong> Early face-encoding response (N170 often occipito-temporal; VPP as its vertex-positive counterpart).</li><li><strong>Clinical pattern (broad):</strong> Systematic review work reports <strong>more consistent N170/VPP reductions to faces in schizophrenia</strong> than in ASD, but it’s not a stand-alone clinical test and is sensitive to stimulus properties (faces vs objects, emotion, gaze). (<a href="https://pubmed.ncbi.nlm.nih.gov/24923618/?utm_source=chatgpt.com">PubMed</a>)</li></ul><h3>9. Oscillatory “circuit probes” (E/I balance, coordination)</h3><h4>40-Hz ASSR (auditory steady-state; power + phase-locking)</h4><ul><li><strong>What it is:</strong> Entrainment to rhythmic auditory stimulation around 40 Hz; readouts include <strong>power</strong> and <strong>phase-locking/ITC</strong>.</li><li><strong>Clinical pattern:</strong> Robust schizophrenia-spectrum deficits are repeatedly reported, making it attractive as a mechanistic “E/I balance / PV-interneuron” probe in translational work. (<a href="https://jamanetwork.com/journals/jamapsychiatry/fullarticle/2566207">jamanetwork.com</a>)</li></ul><h4>Resting frontal alpha asymmetry (FAA)</h4><ul><li><strong>What it is:</strong> A resting EEG asymmetry metric (often F4–F3 in alpha band) historically linked to approach/withdrawal models.</li><li><strong>State of evidence:</strong> Recent meta-analytic summaries indicate <strong>small group effects</strong> and emphasize <strong>limited diagnostic value</strong> plus sensitivity to analytic choices (reference, epoching, alpha definition). (<a href="https://www.semanticscholar.org/paper/Meta-analysis-of-resting-frontal-alpha-asymmetry-as-Luo-Tang/12d28708ebf7bd8c3b8656928963a4dc44bd3e5e?utm_source=chatgpt.com">semanticscholar.org</a>)</li></ul><h3>10. Network dynamics &amp; “whole-brain” summaries</h3><h4>EEG microstates (A–D temporal parameters)</h4><ul><li><strong>What it is:</strong> Recurrent whole-scalp topographies stable for <strong>~60–120 ms</strong> (often summarized as ~90 ms). Metrics include <strong>duration</strong>, <strong>time coverage</strong>, and <strong>occurrence</strong>. (<a href="https://www.nature.com/articles/s41467-020-16914-1">Nature</a>)</li><li><strong>Schizophrenia endophenotype signal:</strong> In an open-access Nature Communications study, <strong>patients and unaffected siblings</strong> showed <strong>increased microstate C</strong> and <strong>decreased microstate D</strong> vs controls, supporting candidacy as an endophenotype. (<a href="https://www.nature.com/articles/s41467-020-16914-1">Nature</a>)</li><li><strong>Interpretation caution:</strong> Links to fMRI resting-state networks are suggestive but not definitive; pipeline choices (number of classes, clustering) matter materially. (<a href="https://www.nature.com/articles/s41467-020-16914-1">Nature</a>)</li></ul><h3>11. Sleep EEG endophenotypes with psychiatric relevance</h3><h4>Sleep spindles (sigma; density/amplitude; NREM Stage 2)</h4><ul><li><strong>What it is:</strong> A defining NREM Stage 2 oscillation implicated in <strong>memory consolidation</strong>, coordinated with other NREM rhythms.</li><li><strong>Schizophrenia signal:</strong> Reviews describe a <strong>specific spindle deficit in patients and unaffected relatives</strong>, correlating with impaired sleep-dependent memory consolidation, positive symptoms, and abnormal thalamocortical connectivity — implicating <strong>thalamic reticular nucleus</strong> dysfunction as a mechanistic node. (<a href="https://pmc.ncbi.nlm.nih.gov/articles/PMC7307009/">pmc.ncbi.nlm.nih.gov</a>)</li></ul><h3><strong>12. Closing chapter: A reference you can actually use</strong></h3><p>If you’ve made it this far, you’ve probably felt the core tension that runs through psychiatric neuroscience: mental health care deals with real suffering, yet many of our tools are still built around symptoms, interviews, and trial-and-error. This marker library exists to add one practical layer of objectivity — not as a shortcut to diagnosis, but as a clearer way to ask better questions. Across sensory filtering, prediction error, attention and working memory, reward learning, and broader “brain state” measures, the common theme is simple: EEG can offer measurable signals that map to functions and circuits we care about. Used well, these markers help us move from vague labels toward testable mechanisms.</p><p>The most important takeaway is also the most grounding: these markers are <em>research- and stratification-grade</em>. They become valuable when you treat them as probabilistic probes and specify the details that make them meaningful — <strong>the paradigm, preprocessing choices, the scoring window, and the scalp topography</strong>. In other words, a marker is not just a name like “MMN” or “P300”; it’s a defined measurement recipe. That is exactly why a library format matters: it turns scattered findings into a shared reference, improves communication across teams, and makes it easier to compare results across studies, cohorts, and platforms. For clinical research, this supports smarter trial design (enrichment, subtyping, endpoint selection). For clinical practice, it points toward a future where monitoring and personalization become more feasible — even if we’re not fully there yet.</p><p>Consider this library a starting point, not a verdict. The next step is application: pick a marker, define it precisely, link it to a hypothesis, and test it in the real world — with transparency about limits, confounders, and reproducibility. If you’re building, investing, or writing, use this as a map: a quick way to orient yourself, identify what’s credible, and decide what deserves deeper investigation. And if it sparks curiosity, that’s the best outcome — because psychiatry will progress fastest when more people can navigate the science with clarity, rigor, and the confidence to explore.</p><h3>🧩 About me</h3><p>With a background spanning Cognitive Neurosciences, <strong>R&amp;D in Molecular Neurobiology</strong>, Clinical Development Management, I am earning my stripes in Venture Capital, because I love to explore new perspectives to evaluate Tech innovation and <strong>investments in Life Sciences</strong>. My commitment to transforming scientific insights into tangible health-tech solutions fuels my passion for this evolving AI-healthcare intersection especially when related to the brain and mental health.</p><p>If you want to discuss these exciting trends further — please feel free to reach out on <a href="https://www.linkedin.com/in/ugo-tomasello/">LinkedIn</a> or via <a href="http://ugo.tomasello@gmail.com/">email</a>.</p><p>👉 <strong>You can find me on </strong><a href="https://www.linkedin.com/in/ugo-tomasello/"><strong>LinkedIn</strong></a><strong> and keep up to date with my latest insights on </strong><a href="https://medium.com/@ugotomasello"><strong>Medium</strong></a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=a2bd720d8aef" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Why Your Next Therapist Might Be an AI — And Why That’s Not a Bad Thing]]></title>
            <link>https://medium.com/included-vc/why-your-next-therapist-might-be-an-ai-and-why-thats-not-a-bad-thing-3684657c4692?source=rss-6bd5914fe82e------2</link>
            <guid isPermaLink="false">https://medium.com/p/3684657c4692</guid>
            <category><![CDATA[venture-capital]]></category>
            <dc:creator><![CDATA[Ugo Tomasello]]></dc:creator>
            <pubDate>Thu, 24 Apr 2025 17:47:09 GMT</pubDate>
            <atom:updated>2025-05-27T11:51:40.665Z</atom:updated>
            <content:encoded><![CDATA[<h3><em>Why Your Next Therapist Might Be an AI — And Why That’s Not a Bad Thing</em></h3><figure><img alt="An illustration of a person holding a smartphone, with a rainbow-colored arc emerging from the screen. The person’s head is depicted with cloud-like shapes and a lightning bolt, symbolizing mental health. Surrounding the figure are logos of mental health and AI companies, including BetterHelp, Modern Health, Woebot, Wysa, Teladoc, Amwell, Ginger, Lyra, Cerebral, Google DeepMind, Limbic, Kooth, Acute XR, Talkspace, and Amazon Clinic." src="https://cdn-images-1.medium.com/max/1024/1*yY_JEAWQo1BI2fFD80fcNA.png" /><figcaption>Cover image</figcaption></figure><p>· <a href="#801f">🛋️ The Digital Couch Is Open: How We Got Here (and What’s Next)</a></p><p>· <a href="#c393">✨ From AI to VR — Technological Advances Transforming Tele-Mental Health</a><br> ∘ <a href="#3287">🔍 AI That Listens, Learns, and Predicts</a><br> ∘ <a href="#32b8">🤖 AI Chatbots: The Digital Therapist That Never Sleeps</a><br> ∘ <a href="#2631">⌚ Your Smartwatch Knows You’re Struggling Before You Do</a><br> ∘ <a href="#f208">🕶️ Virtual Reality Therapy: Healing the Mind in a Digital World</a><br> ∘ <a href="#cf16">Real case: Acute XR, when VR help your kid to control anxiety</a></p><p>· <a href="#2233">🏛️ The Tele-Mental Health Market — From Rapid Adoption to Sustainable Growth</a><br> ∘ <a href="#f345">🚀 Adoption and Uptake: Tele-Mental Health Goes Mainstream</a><br> ∘ <a href="#8f98">🏢 Key Industry Players: From Teletherapy Giants to AI Innovators</a><br> ∘ <a href="#8047">🔹 Early Pioneers and Teletherapy Giants</a><br> ∘ <a href="#83e2">🔹 Employer-Focused Platforms and Hybrid Care Models</a><br> ∘ <a href="#7cea">🔹 AI-Powered Mental Health Chatbots and Self-Guided Tools</a><br> ∘ <a href="#82a5">🔹 Telepsychiatry Platforms and Virtual Medication Management</a><br> ∘ <a href="#361b">🔹 Big Tech’s Entry into AI-Powered Mental Health</a><br> ∘ <a href="#428b">🔹 Mergers, Acquisitions, and Market Consolidation</a><br> ∘ <a href="#9ef4">🔹 New Entrants and Niche Solutions</a><br> ∘ <a href="#0b15">💰 The Investment Landscape: From Pandemic Boom to Evidence-Based Growth</a><br> ∘ <a href="#03d1">Largest Deals and Notable Startups</a><br> ∘ <a href="#c6e2">📉 Sustainability and ROI Expectations from Investors</a><br> ∘ <a href="#f9d7">🔄 Evolving Business Models: From Subscription Fatigue to Hybrid Care</a><br> ∘ <a href="#1aff">🏦 AI as a Cost-Saving and Productivity Tool for Insurers and Employers</a></p><p>· <a href="#027a">🔮 The Future of Digital Mental Health: From Hype to Healing</a></p><p>· <a href="#5f5e">🧩 About me</a></p><h3>🛋️ The Digital Couch Is Open: How We Got Here (and What’s Next)</h3><p>What if your next therapist isn’t a human sitting across from you, but an AI chatbot that knows exactly what to say? It’s not science fiction anymore (<em>Marr, 2023</em>). Over the past few years, telemedicine has quietly transformed how we access mental health care. Once a niche service, digital therapy went mainstream almost overnight during the COVID-19 pandemic (<em>Cureus, 2023</em>). Video calls with therapists, once rare, became a lifeline for millions coping with anxiety, isolation, and grief (<em>DLA Piper, 2023</em>). In the U.S. alone, virtual mental health visits surged from less than 1% of outpatient visits pre-pandemic to over 40% at its peak (<em>Cureus, 2023</em>). Even today, tele-mental health still accounts for approximately <strong>36% of outpatient mental health care</strong> in the United States (<em>Cureus, 2023</em>).</p><p>Suddenly, geography no longer mattered. Therapy was in your living room, on your lunch break, or available at midnight. But this was just the beginning. What started as Zoom calls with your therapist has evolved into something far more ambitious. Artificial intelligence, large language models, and virtual reality are rapidly changing the rules of mental health care. AI-powered chatbots like <strong>Woebot</strong> and <strong>Wysa</strong> are already coaching users through cognitive behavioral therapy (<strong>CBT</strong>), providing 24/7 support that millions have turned to in moments of need (<em>Snoswell et al., 2023; Zhang et al., 2024; Marr, 2023</em>).</p><p>Remote patient monitoring (<strong>RPM</strong>) apps are quietly tracking mood swings and behavior in real time, helping clinicians detect warning signs early (<em>Cureus, 2023; DLA Piper, 2023</em>). VR exposure therapy is helping patients confront phobias and PTSD triggers — without ever leaving home (<em>Zhang et al., 2024; Marr, 2023</em>). By 2025, tele-mental health is no longer a stopgap solution; it’s a permanent pillar of global mental health services. <strong>Surveys show that 94–99% of patients report being “very satisfied”</strong> with their telehealth experiences (<em>DLA Piper, 2023; Cureus, 2023</em>). Meanwhile, 88% of U.S. physicians now offer telehealth services, up from just 43% pre-pandemic (<em>DLA Piper, 2023</em>).</p><p>For many, <strong>virtual therapy isn’t second best</strong> — it’s their first choice. And as AI takes center stage, the line between human-guided and machine-supported therapy is starting to blur (<em>Snoswell et al., 2023; Marr, 2023; Zhang et al., 2024</em>).</p><p>But that was only the beginning.</p><figure><img alt="A graphic comparing RPM (Remote Patient Monitoring) and CBT (Cognitive Behavioral Therapy) in telehealth. It highlights 24/7 chatbot support for CBT, and that virtual mental health visits rose from 1% to 40%. For RPM, it notes 94–99% of patients are “very satisfied,” and that physicians offering telehealth increased from 43% pre-pandemic to 88% post-pandemic. Arrows indicate interaction between RPM and CBT." src="https://cdn-images-1.medium.com/max/1024/1*KpErH--ktTcQJJTm7e9wsA.png" /><figcaption>Remote Patient Monitoring and Cognitive Behavioral Therapy</figcaption></figure><h3>✨ From AI to VR — Technological Advances Transforming Tele-Mental Health</h3><p>Tele-mental health has entered a new phase defined by <strong>AI-enhanced diagnostics, chatbot-based therapy, wearable-driven remote monitoring, and immersive virtual reality treatment environments</strong>. Each of these technologies addresses critical gaps in mental health access, personalization, and scalability.</p><h4>🔍 AI That Listens, Learns, and Predicts</h4><p>Imagine if your therapist could detect your stress levels just by listening to your voice or tracking your sleep patterns. Thanks to artificial intelligence, that future is already here. AI-powered tools can now analyze <strong>speech patterns, facial expressions, and even passive behavioral data</strong> to assess a person’s mental state <strong>with increasing accuracy</strong> (<em>Zhang et al., 2024​; WHO, 2023</em>​).</p><p>For instance, <strong>new AI models are being trained to predict cognitive decline</strong> years before symptoms appear by analyzing EEG patterns during sleep (<em>New York Post, 2025</em>​). Other algorithms can scan patient speech for <strong>subtle vocal markers of depression and anxiety</strong>, often detecting signs that even experienced clinicians might miss (<em>Coded Companions Report, 2024</em>​).</p><p>But while AI’s ability to detect risk is promising, it also raises <strong>complex ethical challenges</strong>. Researchers warn that <strong>algorithmic bias remains a major issue</strong>, as AI models tend to perform worse for marginalized communities if their training data is not diverse (<em>WHO, 2023​; JMIR Mental Health, 2024</em>​). Moreover, <strong>clinicians remain skeptical about over-reliance on AI</strong>, with studies showing that <strong>lack of transparency</strong> in how these models make decisions leads to hesitation in adoption (<em>JMIR Formative Research, 2024​</em>).</p><p>For now, AI works best as a <strong>co-pilot</strong> for human clinicians — providing insights, flagging risks, and supporting early interventions rather than making final decisions. Transparency and human oversight are essential to ensure AI tools augment, rather than replace, human clinical judgment (<em>WHO, 2023</em>).</p><figure><img alt="A quadrant diagram titled “Balancing AI in Mental Health Care.” It shows four areas: 1) Algorithmic Bias — low effectiveness with high ethical concerns; 2) AI as Co-Pilot — balanced effectiveness and ethical oversight; 3) Skepticism of Over-Reliance — low ethical concerns but high distrust; and 4) Early Intervention Tools — high effectiveness with minimal ethical issues. Each section uses different colors and icons." src="https://cdn-images-1.medium.com/max/1024/1*yIOXZRhu3nB177ycbNEamw.png" /></figure><h4>🤖 AI Chatbots: The Digital Therapist That Never Sleeps</h4><p>Not everyone has immediate access to a human therapist. That’s where AI-powered mental health chatbots come in. Apps like <strong>Woebot, Wysa, and Limbic</strong> are offering <strong>real-time, conversational support</strong> to millions, coaching users through <strong>CBT</strong> techniques and mood-tracking exercises (<em>AI Chatbots for Mental Health Review, 2024</em>​).</p><p>These bots are not just a gimmick — they have <strong>demonstrated clinical effectiveness</strong>. A large-scale study found that AI-driven conversational agents can <strong>reduce symptoms of anxiety and depression to a degree comparable to human-guided interventions</strong> (<em>JMIR Mental Health, 2024​</em>). Another study suggests that <strong>chatbots may help bridge the gap for people reluctant to seek therapy</strong>, as users report feeling <strong>less judged when speaking to AI</strong> rather than a human therapist (<em>It Happened to Be the Perfect Thing, 2024</em>​).</p><p>However, chatbot therapy has limitations. <strong>Engagement remains a challenge</strong>, as many users abandon chatbot interactions after a few sessions, particularly if responses feel too generic or repetitive (<em>JMIR Preprint, 2024​</em>). <strong>Ethical concerns</strong> have also emerged — some chatbots, including mental health AI tools deployed in school settings, <strong>have been criticized for collecting sensitive data with insufficient transparency</strong> (<em>WSJ, 2024​</em>).</p><p><strong>AI chatbots should not replace human therapists</strong> but instead serve as an <strong>accessible first step</strong> — a <strong>“bridge”</strong> that provides immediate support until professional help is available (<em>Large Language Models in Healthcare, 2024</em>​).</p><h4>⌚ Your Smartwatch Knows You’re Struggling Before You Do</h4><p>Beyond AI chatbots, wearables and smartphone apps are quietly revolutionizing mental health care. The emerging field of <strong>digital phenotyping</strong> is making it possible to <strong>monitor a person’s mental health based on their digital behavior</strong> — everything from <strong>GPS movements and typing speed to heart rate and sleep patterns</strong> (<em>Wikipedia, 2024​; JMIR Mental Health, 2023​</em>).</p><p>For example, <strong>a smartwatch that detects several nights of poor sleep and increased heart rate variability could flag early signs of a depressive episode</strong> (<em>Resp-BoostNet Study, 2024</em>​). Researchers are also developing AI-powered tools that can <strong>detect rising stress levels in real-time</strong>, prompting personalized coping strategies before symptoms escalate (<em>Digital Trends, 2024</em>​). Platforms like Kooth integrate RPM data to personalize interventions and alert clinicians to early warning signs of crisis (<em>Kooth Economic Evaluation, 2024</em>).</p><p>But <strong>where do we draw the line?</strong> The potential for <strong>over-surveillance</strong> is a growing concern. If AI can predict stress or emotional distress, <strong>who has access to this information?</strong> Employers? Insurance companies? Researchers argue that stronger privacy safeguards are needed, ensuring that users remain in control of their data (<em>Healthcare Access Denied Report, 2024</em>). These concerns are not well addressed yet, and where we favor the medical results, privacy and ethics might be at risk. Continuous monitoring could be a <strong>game-changer</strong> — helping clinicians <strong>intervene earlier, rather than waiting for patients to reach a crisis point</strong> (<em>Telehealth and Telemedicine Handbook, 2024</em>) and probably the line should be adapted for each patient, disease and country, to gather the most intended outcome each time.</p><h4>🕶️ Virtual Reality Therapy: Healing the Mind in a Digital World</h4><p>Perhaps the most sci-fi of all the new mental health technologies is <strong>Virtual Reality Therapy (VR-T)</strong>. By immersing patients in computer-generated environments, VR therapy allows people to confront fears, process trauma, and practice social interactions in a safe, controlled space (<em>Medical News Today, 2024</em>​).</p><p>In California, VR therapy is being used to reduce aggression and improve <strong>emotional regulation</strong> for inmates in solitary confinement (<em>The Guardian, 2025</em>). Meanwhile, in the UK, VR is <strong>helping patients with cocaine addiction</strong> manage cravings by exposing them to simulated high-risk scenarios — providing a safe environment to practice coping strategies without real-world consequences (<em>The Guardian, 2024</em>).</p><p>Research suggests that VR exposure therapy is just as effective as traditional methods for treating PTSD and phobias — and sometimes even more so, thanks to its immersive nature, which increases patient engagement and emotional processing (<em>Medical News Today, 2024</em>​).</p><p>However, <strong>accessibility remains a challenge</strong>. High-end VR equipment is still expensive, and many therapists lack training in how to use VR therapeutically (<em>Social Prescribing in the Metaverse, 2024</em>). But as hardware costs decrease and adoption rises, VR therapy is likely to become a standard option in tele-mental health care.</p><figure><img alt="A quadrant diagram titled “Mapping AI in Mental Health Care.” It highlights four areas: 1) Digital Phenotyping — monitors mental health but raises ethical issues; 2) AI Chatbots — offer support but pose privacy concerns; 3) VR Therapy — less accessible with minimal ethical concerns; and 4) AI-Powered Tools for Clinicians — enhance insights with minimal dilemmas. Each area is color-coded with icons." src="https://cdn-images-1.medium.com/max/1024/1*ytfy0dFtic1tYBu18bbzuQ.png" /><figcaption>Mapping AI in Mental Health Care</figcaption></figure><h4>Real case: Acute XR, when VR help your kid to control anxiety</h4><figure><img alt="Acute XR logo" src="https://cdn-images-1.medium.com/max/1024/0*AY1n6jmcEof1KldL.png" /><figcaption><a href="https://acutexr.com/">https://acutexr.com/</a></figcaption></figure><p>To further understand how mental health technology is opening the venue to new interventions and options for patients, I interviewed with <a href="https://www.linkedin.com/in/dfkirchdorfer?miniProfileUrn=urn%3Ali%3Afs_miniProfile%3AACoAABE6vzoBAyrjTw0YuUBUeS_P0jMVXbjokn8&amp;lipi=urn%3Ali%3Apage%3Ad_flagship3_search_srp_all%3BPOywGJHDRQepFEr35OYK%2FA%3D%3D">Dominik F. Kirchdorfer</a>, CEO of Acute XR, a startup empowering pediatric patients and therapists through play.</p><p><strong><em>Can you explain what is Acute XR and its mission?</em></strong></p><p><strong>Dominik</strong>: “Acute XR is a digital health gaming company that specializes in immersive VR tools for pediatric mental health. While our technology has broader applications and we see great demand from playful adults, our current focus is on pediatrics. We’ve developed a game called Calm Quest, which integrates biofeedback to support therapy. Instead of using traditional controls, patients wear a heart rate sensor — such as a fitness tracker or smartwatch — which acts as the primary input. The game responds to their physiological signals, allowing them to interact by regulating their own body responses.</p><p>Although we call it a game, it’s more of a structured immersive experience. It’s designed to assist therapists in helping children regulate their emotions, improving session efficiency, and enhancing patient engagement. Many therapists struggle with keeping young patients engaged, which can consume a significant portion of their sessions. By integrating play into therapy, we make the process more enjoyable, ultimately increasing the likelihood that children will continue treatment and see real benefits.”</p><p><strong><em>How does Calm Quest work in practise?</em></strong></p><p><strong>Dominik</strong>: “The patient wears a VR headset, and the therapist monitors the experience on a connected device, such as a laptop, tablet, or smartphone. Our platform is designed to be highly accessible, eliminating the need for powerful gaming PCs while maintaining a high-quality visual experience.</p><p>Therapists can adjust settings in real time, such as difficulty levels, based on the patient’s reactions. The system also includes tools for therapists to track progress, such as self-assessment prompts and heart rate variability data, allowing them to analyze trends over multiple sessions. The key is that therapy remains therapist-guided — patients are encouraged to relax or adjust behaviors in response to in-game feedback, reinforcing emotional regulation skills.”</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fwww.youtube.com%2Fembed%2FLg4uSQ6VROQ%3Ffeature%3Doembed&amp;display_name=YouTube&amp;url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DLg4uSQ6VROQ&amp;image=https%3A%2F%2Fi.ytimg.com%2Fvi%2FLg4uSQ6VROQ%2Fhqdefault.jpg&amp;type=text%2Fhtml&amp;schema=youtube" width="854" height="480" frameborder="0" scrolling="no"><a href="https://medium.com/media/a95f3a1c0ce6876b80f21be8a475c0d4/href">https://medium.com/media/a95f3a1c0ce6876b80f21be8a475c0d4/href</a></iframe><p><strong><em>Why VR ? what is the added value of Acute XR?</em></strong></p><p><strong>Dominik</strong>: “We conducted a clinical trial at Vienna General Hospital, which confirmed that our approach is effective. However, the most groundbreaking insight was the immediate immersion effect<strong>. Unlike traditional therapeutic methods, where it can take 20 or even 30 minutes for a child to actively engage, VR captures their attention instantly.</strong></p><p>This is a major advantage, especially in pediatric therapy. Young children often struggle to understand why they are in therapy, and teenagers may be reluctant to discuss their emotions. Therapists frequently lose a significant portion of their sessions simply trying to establish a connection. With Calm Quest, this challenge disappears. From the moment the headset is on, kids are engaged — asking how to play, interacting with the environment, and naturally participating in the session. The efficiency gains for therapists are immense, reducing wasted time and increasing the effectiveness of each session.”</p><p><strong><em>Currently, Acute XR is only available for therapists. Why not a home version?</em></strong></p><p><strong>Dominik</strong>: “A home version is planned, primarily because therapists have requested a ‘homework mode.’ Just like traditional therapy, where patients are assigned exercises between sessions, VR therapy can also benefit from structured at-home practice. <strong>Therapists have told us this is the biggest limitation in VR therapy right now.</strong></p><p><strong>However, we don’t believe therapy should be fully automated or self-guided.</strong> The role of the therapist is crucial — they tailor interventions based on individual needs, ensuring the exercises are effective. While Calm Quest provides immersive biofeedback-driven experiences, it doesn’t replace clinical expertise. It is a tool, not a therapy. For instance, in the game, a misty forest brightens and transforms as the patient calms down. <strong>This visual feedback is powerful, but without proper guidance, a patient may not know how to achieve or sustain that state.</strong></p><p>The planned home version will prioritize entertainment while also allowing therapists to assign exercises remotely. This ensures continuity in therapy without compromising its effectiveness.”</p><p><strong><em>Which are the next steps for Acute XR?</em></strong></p><p><strong>Dominik</strong>: “We have ambitious plans. <strong>We’re developing a clinical version of our platform that integrates AI, working with IBM’s WatsonX AI to enhance therapeutic applications.</strong> This version is being designed for hospitals and psychiatric care, requiring regulatory approvals from entities like the EMA and FDA.</p><p>We’re exploring various applications that could address specific pain points for psychiatrists and hospital management. One such use case is for PTSD treatment. The U.S. military is already using VR for exposure therapy, recreating realistic combat scenarios. We envision similar applications for civilians, such as trauma survivors from abuse, accidents, or other distressing events. AI-generated environments could personalize exposure therapy, enabling patients to process trauma in a controlled, therapeutic setting. This is just one of many ways we see VR revolutionizing mental health care. I want to stress once more that our aim is not to replace doctors or therapists, but aid them in their work<strong>. We believe that A.I. has huge potential for the healthcare sector, but A.I. should never replace people. It should make their jobs easier and improve their outcomes for the betterment of all.</strong>”</p><blockquote>Because they’re kids and they want to just have fun. They want to play.</blockquote><h3>🏛️ The Tele-Mental Health Market — From Rapid Adoption to Sustainable Growth</h3><p>Tele-mental health and AI-driven mental health solutions have evolved from pandemic-era lifelines into a permanent pillar of healthcare delivery. What started as a scramble to maintain services during lockdowns has become a maturing global industry, reshaping patient expectations, clinician workflows, and business models alike.</p><h4>🚀 Adoption and Uptake: Tele-Mental Health Goes Mainstream</h4><p>By 2022, 96% of large employers in the U.S. were offering telehealth services — including mental health care — as part of their benefits packages, a remarkable increase from just 46% in 2019 (<em>SmithBrothersUSA, 2023</em>). Insurers, including private health plans and Medicaid programs, have permanently expanded coverage for tele-mental health after witnessing its efficacy during the pandemic (<em>KFF, 2023</em>).</p><p>This widespread adoption has extended beyond urban centers. In 2021, 55% of outpatient mental health visits in rural areas were delivered via telehealth — compared to 35% in urban areas — demonstrating tele-mental health’s power to close gaps in access (<em>KFF, 2023</em>). Meanwhile, convenience continues to drive demand: surveys report that 57–80% of patients prefer telehealth over in-person appointments due to time savings and the comfort of receiving care at home (<em>TechTarget, 2024</em>).</p><h4>🏢 Key Industry Players: From Teletherapy Giants to AI Innovators</h4><h4>🔹 Early Pioneers and Teletherapy Giants</h4><p>Companies like <strong>BetterHelp (</strong>owned by Teladoc Health) and <strong>Talkspace</strong> were among the first to disrupt traditional therapy by offering direct-to-consumer, subscription-based access to licensed therapists. Both platforms boast large networks — BetterHelp claims over 30,000 licensed therapists on its platform — offering users options to match with professionals tailored to their needs (<em>Enhancing Mental Health with AI, 2024</em>).</p><p>BetterHelp and Talkspace pioneered asynchronous messaging therapy models, providing unlimited chat or text-based therapy under fixed monthly fees. These offerings dramatically reduced wait times, allowing users to connect with a therapist within days, compared to the weeks or months typical of traditional therapy settings (<em>SmithBrothersUSA, 2023</em>).</p><p>However, these platforms have faced scrutiny for issues like therapist burnout, inconsistent quality of care, and data privacy concerns. BetterHelp, for example, faced an FTC fine in 2023 over allegations of sharing sensitive user data with advertisers (<em>FTC.gov, 2023</em>).</p><h4>🔹 Employer-Focused Platforms and Hybrid Care Models</h4><p><strong>Lyra Health</strong> and <strong>Modern Health</strong> have emerged as leading players in the employer-sponsored mental health space. These platforms focus on enterprise clients, providing comprehensive mental health benefits that integrate digital self-care tools with live, one-on-one therapy and coaching sessions.</p><p>Lyra Health uses AI algorithms to match employees with therapists and coaches, improving engagement and outcomes by ensuring better fit and availability (<em>Enhancing Mental Health with AI, 2024</em>). Modern Health offers tiered levels of care, including digital CBT programs, meditation modules, and personalized coaching, complementing live therapy for employees at companies like Pixar, Zendesk, and Clif Bar.</p><p>These hybrid models are gaining traction among employers seeking scalable, cost-effective mental health solutions that also reduce absenteeism and improve workforce productivity (<em>AI and Mental Healthcare Opportunities, 2024</em>).</p><h4>🔹 AI-Powered Mental Health Chatbots and Self-Guided Tools</h4><p><strong>Woebot</strong> and <strong>Wysa</strong> are two of the most widely used AI-driven mental health chatbots offering 24/7 support. Both leverage CBT techniques, delivering psychoeducation, mood tracking, and mindfulness exercises via mobile apps.</p><p>Woebot was developed by clinical psychologists at Stanford and has been validated in multiple studies for reducing symptoms of depression and anxiety (<em>AI Chatbots for Mental Health Review, 2024</em>). Wysa, an AI-powered “emotionally intelligent” chatbot, has served over 5 million users globally and is employed by health insurers like Aetna for workplace mental health support (<em>Enhancing Mental Health with AI, 2024</em>).</p><p>Despite their promise, these tools are designed as supplements rather than replacements for human-led therapy. Without appropriate oversight, there are concerns about the accuracy of AI-generated advice and the potential risks in crisis situations (<em>Building Trust in Artificial Intelligence and New Technologies in Mental Health, 2024</em>).</p><h4>🔹 Telepsychiatry Platforms and Virtual Medication Management</h4><p>Platforms like <strong>Cerebral</strong> and <strong>Brightside Health</strong> focus on delivering psychiatric consultations and medication management services. Their AI-driven intake assessments streamline diagnosis and treatment planning, often offering same-day virtual appointments for medication prescriptions.</p><p>Brightside uses machine learning to predict how patients will respond to different medications, optimizing pharmacological treatment plans (<em>Cureus AI in Telehealth, 2024</em>). However, Cerebral has faced intense regulatory scrutiny for allegedly overprescribing controlled substances, prompting investigations by the Department of Justice and DEA in 2022 (<em>Cureus AI in Telehealth, 2024</em>).</p><p>These controversies highlight the delicate balance between innovation and ethical clinical practice in AI-driven psychiatry.</p><h4>🔹 Big Tech’s Entry into AI-Powered Mental Health</h4><p>Big Tech companies are increasingly investing in mental health services, leveraging their vast resources and AI expertise.</p><ul><li><strong>Amazon Clinic</strong> has expanded its telehealth services to include virtual mental health consultations. While its focus has initially been on primary care, Amazon’s acquisition of One Medical signals deeper intentions to integrate behavioral health into its virtual care offerings (<em>Building Robust Regulation of Digital Mental Health, 2024</em>).</li><li><strong>Google’s DeepMind</strong> is developing AI models to enhance psychological assessment capabilities, with ongoing research into large language models that could support clinical decision-making in mental health (<em>Building Robust Regulation of Digital Mental Health, 2024</em>).</li><li><strong>Apple</strong>, through its Health app and ecosystem of wearables, is quietly building tools to track mental health markers, including mood logging and sleep tracking, positioning itself to offer more proactive mental wellness interventions in the future (<em>Enhancing Mental Health with AI, 2024</em>).</li></ul><h4>🔹 Mergers, Acquisitions, and Market Consolidation</h4><p>As competition intensifies, mergers and acquisitions are reshaping the tele-mental health landscape.</p><p>The 2021 merger of <strong>Headspace</strong> (meditation app) and <strong>Ginger</strong> (teletherapy provider) to form <strong>Headspace Health</strong> marked the convergence of self-help tools and clinical care into a single platform. This vertical integration strategy aims to offer end-to-end mental health services, from self-care to therapy and psychiatry (<em>Enhancing Mental Health with AI, 2024</em>).</p><p>At the same time, telehealth giants like Teladoc and Amwell continue to acquire niche mental health startups to bolster their capabilities, often adding AI-powered diagnostic tools and therapist-matching algorithms to their core platforms (<em>Galengrowth, 2024</em>).</p><h4>🔹 New Entrants and Niche Solutions</h4><p>Innovative startups are addressing niche markets often overlooked by traditional providers.</p><ul><li><strong>Brightline</strong>, for instance, specializes in pediatric behavioral health, offering virtual therapy and coaching for children and teens.</li><li><strong>Spring Health</strong> combines AI-driven precision mental health care with culturally competent therapy options, addressing the growing demand for inclusive mental health services (<em>AI and Mental Healthcare Opportunities, 2024</em>).</li></ul><h4>💰 The Investment Landscape: From Pandemic Boom to Evidence-Based Growth</h4><p>Venture capital funding for digital mental health skyrocketed between 2019 and 2021. In 2021 alone, mental health startups raised $7.4 billion globally — making it the highest-funded therapeutic category that year (<em>Economic Evaluation of 27,540 Patients, 2024</em>).</p><p>However, the landscape shifted by 2023. Investor enthusiasm cooled as funders demanded stronger evidence of clinical efficacy, regulatory compliance, and sustainable unit economics before committing capital (<em>Building Robust Regulation of Digital Mental Health, 2024</em>). <strong>Quarterly trends</strong> in 2023 showed minimal growth, with <strong>total U.S. digital health VC funding</strong> dropping from <strong>$15.3 billion in 2022 to $10.7 billion in 2023 </strong>(Rock Health H1 2024 Report). Startups that had scaled quickly during the pandemic struggled with customer retention and reimbursement complexities. After two years of contraction, <strong>global funding for digital mental health rebounded in 2024</strong>, reaching <strong>$2.7 billion across 184 deals</strong>, marking a <strong>38% increase from 2023</strong>. Digital mental health now represents <strong>12% of all digital health funding</strong>, second only to oncology (<a href="https://galengrowth.com/">Galen Growth, 2024</a>).</p><p>Despite the slowdown, investor interest in AI-driven mental health platforms remains strong. Platforms that integrate AI into hybrid care models — blending automated triage and monitoring with human therapy — are attracting significant funding. Investors now prioritize companies demonstrating measurable outcomes, cost savings, and long-term value (<em>Enhancing Mental Health with AI, 2024</em>).</p><p>For example, one economic evaluation found that the Kooth platform, a web-based mental health service for children and young people, reduced healthcare costs by 40% while maintaining high levels of user satisfaction (<em>An Early Economic Evaluation of Kooth, 2024</em>).</p><figure><img alt="A stacked bar chart showing venture capital deployed to digital health ventures from 2020 to 2024, highlighting mental health funding. Mental health funding peaked at $7.4B in 2021, dipped to $1.9B in 2023, and is projected to rise 38% to $2.7B in 2024. Global digital health funding (excluding mental health) peaked at $53.1B in 2021 and declined to $21.9B by 2024." src="https://cdn-images-1.medium.com/max/1024/1*Ep4e_31JebGFCampZTRCcg.png" /><figcaption><a href="https://www.galengrowth.com/mental-healths-investment-resurgence-a-market-ripe-for-innovation/">VC investment decline across years in mental health</a></figcaption></figure><h4>Largest Deals and Notable Startups</h4><p>Investors are shifting <em>where</em> they deploy capital in this space. There is a tilt toward <strong>late-stage funding</strong> — in fact, 2024 saw the highest level of late-stage (Series C and beyond) funding for mental health since 2021. ​Rather than placing many risky bets on brand-new startups, VCs are doubling down on companies that have proven models and solid revenue.</p><ul><li><strong>Talkiatry</strong> raised a <strong>$130 million Series C</strong> in mid-2024 to scale its hybrid psychiatry services across the U.S. (PR Newswire, 2024).</li><li><strong>Spring Health</strong>, offering employer-focused mental health services, extended its <strong>Series C with $71 million</strong> in early 2024, bringing its total funding to over <strong>$370 million</strong> (Business Wire, 2024).</li><li><strong>Brightline</strong>, specializing in pediatric digital mental health, continues to build on its <strong>$105 million Series C</strong> (2022) with strategic partnerships in 2023–2024, especially with payers and Medicaid plans (Brightline Press Release, 2023).</li><li><strong>Koa Health</strong>, based in Spain, secured <strong>€30 million ($36 million)</strong> in 2023, focusing on digital-first mental health solutions for employers and health systems (Tech.eu, 2023).</li><li><strong>Unmind</strong>, a UK-based workplace mental health platform, raised a <strong>$50 million Series B</strong> in mid-2023 (Unmind Newsroom, 2023).</li></ul><figure><img alt="Market map representing digital mental health and AI companies. It includes Woebot, Wysa, Google DeepMind, Limbic, BetterHelp, Talkspace, Cerebral, Koa Health, Brightside Health, Ginger, Spring Health, Unmind, Lyra, Modern Health, Brightline, Kooth, Apple Clinic, Amazon Clinic, Amwell, Teladoc Health, Akili, Acute XR, and Headspace. The layout groups companies by categories or services within mental health tech." src="https://cdn-images-1.medium.com/max/1024/1*aojnbowLBPEy_Z-ZpLB3gQ.png" /><figcaption>Market Map Mental Health Verticals</figcaption></figure><h4>📉 Sustainability and ROI Expectations from Investors</h4><p>For investors, the focus now is on <strong>sustainable growth and outcomes</strong>. They are asking: Do these tele-mental health platforms actually improve patient well-being? Can they reduce healthcare costs (for example, by preventing expensive ER visits)? Are users sticking with the programs? Startups that can demonstrate clinical efficacy — backed by studies or at least robust internal data — have a better shot at premium valuations.</p><ul><li><strong>Clinical Outcomes as a Must-Have</strong>: Investors are prioritizing platforms with <strong>validated clinical efficacy</strong>. For instance, <strong>Kooth’s</strong> U.S. pilot programs demonstrated a <strong>40% reduction in clinical symptom severity</strong> and <strong>30% lower healthcare costs</strong> (Kooth Economic Impact Study, 2024).</li><li><strong>Cost Containment and Unit Economics</strong>: B2B contracts with <strong>per member per month (PMPM)</strong> pricing, as seen with <strong>Spring Health</strong> and <strong>Lyra</strong>, provide <strong>predictable revenue</strong> and <strong>high customer lifetime value</strong> (CB Insights Digital Health Report, 2024).</li><li><strong>Engagement and Retention</strong>: Platforms that demonstrate <strong>high user retention</strong> — especially in <strong>employer mental health programs</strong> — are securing funding. Lyra reports <strong>80% completion rates</strong> for therapy programs, a major selling point for clients and investors (Lyra Health Outcomes Report, 2023).</li><li><strong>Regulatory and Payer Integration</strong>: Solutions integrated into public and private payer networks — such as <strong>Kooth in the UK NHS and California Medi-Cal system</strong> — are viewed as <strong>de-risked investments</strong> (Kooth, 2023).</li></ul><figure><img alt="A bar and line graph showing venture capital (VC) investment in mental health ventures and the share raised by AI-driven startups from 2020 to 2024. VC investment peaked at $4.3B in 2021, then declined. AI ventures accounted for 60% of funding in 2020, dipped to 35% in 2022, and rose to 48% by 2024. Light blue bars represent total VC; a peach line tracks AI’s share." src="https://cdn-images-1.medium.com/max/1024/1*LPe0MmCKYNTr5puhCeHK-w.png" /><figcaption><a href="https://www.galengrowth.com/mental-healths-investment-resurgence-a-market-ripe-for-innovation/">VC investments in AI-driven mental health</a></figcaption></figure><h4>🔄 Evolving Business Models: From Subscription Fatigue to Hybrid Care</h4><p>Early digital therapy platforms like BetterHelp and Talkspace built their businesses on flat-fee, direct-to-consumer subscription models. These models prioritized growth, offering unlimited messaging or weekly video therapy for a monthly fee. Yet, concerns about therapist burnout, retention, and customer churn have led many platforms to evolve (<em>Enhancing Mental Health with AI, 2024</em>). Now, we see models evolving towards sustainability. Some platforms have raised prices or introduced <strong>tiered subscription plans</strong> (e.g., paying more for live video access vs. text-only coaching). Others are integrating with insurance networks so that patients pay regular co-pays and sessions are reimbursed like traditional therapy — a move toward parity with in-person care.</p><p>Today, <strong>hybrid care models</strong> are emerging as the industry standard. Platforms like Lyra Health and Modern Health combine AI-supported tools — such as automated assessments, mood tracking, and chatbots — with live therapy sessions delivered by licensed professionals (Combining AI and Human Support in Mental Health, 2024). Patients prefer these blended models, where AI provides ongoing support and self-help resources between human-led sessions (<em>AI and Mental Healthcare Opportunities, 2024</em>).</p><p><strong>B2B2C partnerships, </strong>tele-mental health firms teaming with insurers, employers, or schools to reach users, are also gaining traction. For instance, colleges might contract a telehealth service to offer counseling for students, or an insurer might bundle a therapy app as a member perk.</p><p>This offer reimbursable services that mirror traditional healthcare coverage (<em>Economic Evaluation of 27,540 Patients, 2024</em>). and helps to reduce out-of-pocket costs for patients and promoting long-term engagement.</p><h4>🏦 AI as a Cost-Saving and Productivity Tool for Insurers and Employers</h4><p>AI-driven platforms appeal to payers and employers seeking to manage healthcare costs and boost employee productivity. AI mental health tools have been shown to reduce hospital admissions related to mental health crises by up to 30% (<em>An Early Economic Evaluation of Kooth, 2024</em>).</p><p>For employers, these tools offer scalable mental health support programs, such as stress management and burnout prevention. Corporations now see mental health investments as essential to maintaining workforce well-being and reducing absenteeism (<em>Enhancing Mental Health with AI, 2024</em>).</p><h3>🔮 The Future of Digital Mental Health: From Hype to Healing</h3><p>The era of digital mental health has matured. What began as a crisis-driven pivot during the pandemic has now evolved into a permanent and powerful pillar of global healthcare. The convergence of AI, telehealth, wearables, and VR is rewriting the rules of mental health delivery — transforming therapy from a one-size-fits-all model into something more scalable, personalized, and accessible.</p><p>But the stakes are higher now. Investors, clinicians, and patients are no longer dazzled by flashy apps or sky-high user acquisition numbers. Today’s digital mental health landscape demands <strong>proof</strong>: proof that these solutions deliver real clinical outcomes, proof that they safeguard patient privacy, and proof that they can scale sustainably without compromising care.</p><p>The winning formula in 2025 and beyond isn’t about replacing human therapists with algorithms. It’s about <strong>empowering clinicians</strong>, <strong>extending access</strong> to underserved populations, and <strong>augmenting human care with AI-driven tools</strong> that enhance outcomes while reducing costs. Hybrid care models — blending digital self-help tools with live therapy — are becoming the industry standard. Companies like <strong>Talkiatry</strong>, <strong>Spring Health</strong>, <strong>Lyra Health</strong>, and <strong>Kooth</strong> exemplify this shift, combining cutting-edge tech with evidence-based clinical frameworks.</p><p>At the same time, ethical questions loom large. As AI-powered chatbots, remote patient monitoring, and VR therapy become ubiquitous, so too do concerns over data privacy, algorithmic bias, and the delicate balance between automation and empathy. Regulators are stepping in, and platforms that proactively embrace transparency, equity, and safety will be better positioned to lead.</p><p>From <strong>AI chatbots that listen 24/7</strong>, to <strong>wearables that detect stress before we feel it</strong>, to <strong>VR worlds that teach us how to manage anxiety</strong>, the future of mental health care is already here — and it’s digital. But it’s not tech for tech’s sake. It’s tech with a purpose: to make mental health care <strong>more human</strong>, <strong>more effective</strong>, and <strong>more accessible</strong> to everyone, everywhere.</p><p>The next frontier? <strong>Personalized, preventative mental health care at scale.</strong> And for those who get it right, the opportunity is massive — both in impact and investment return.</p><h3>🧩 About me</h3><p>With a background spanning <strong>R&amp;D in Molecular Neurobiology</strong>, Clinical Development Management, I am earning my stripes in Venture Capital, because I love to explore new perspectives to evaluate Tech innovation and <strong>investments in Life Sciences</strong>. My commitment to transforming scientific insights into tangible health-tech solutions fuels my passion for this evolving AI-healthcare intersection.</p><p>If you want to discuss these exciting trends further — please feel free to reach out on <a href="https://www.linkedin.com/in/ugo-tomasello/">LinkedIn</a> or via <a href="http://ugo.tomasello@gmail.com/">email</a>.</p><p>👉 <strong>You can find me on </strong><a href="https://www.linkedin.com/in/ugo-tomasello/"><strong>LinkedIn</strong></a><strong> and keep up to date with my latest insights on </strong><a href="https://medium.com/@ugotomasello"><strong>Medium</strong></a></p><p>👉🏼 <strong><em>Follow </em></strong><a href="https://included.vc/"><strong><em>Included VC</em></strong></a><strong><em> on </em></strong><a href="https://twitter.com/included_vc"><strong><em>Twitter</em></strong></a><strong><em> &amp; </em></strong><a href="https://www.linkedin.com/company/includedvc/"><strong><em>LinkedIn</em></strong></a></p><p><strong>#includedvcmafia</strong></p><p><strong>. . .</strong></p><p>🙏 <strong>Special thanks</strong></p><p>A special thank you note to everyone who contributed to the development of this article (the Included VC team, Included VC Delegates, and all the VCs whose work and testimonies allowed me to develop extensive research on the topic). Namely, I would like to thank <a href="https://www.linkedin.com/in/dfkirchdorfer?miniProfileUrn=urn%3Ali%3Afs_miniProfile%3AACoAABE6vzoBAyrjTw0YuUBUeS_P0jMVXbjokn8&amp;lipi=urn%3Ali%3Apage%3Ad_flagship3_search_srp_all%3BPOywGJHDRQepFEr35OYK%2FA%3D%3D">Dominik F. Kirchdorfer</a> and <a href="https://www.linkedin.com/in/nikitathakrar/">Nikita Thakrar</a>.</p><p><strong>References</strong></p><ol><li>Marr, B. (2023). <em>AI In Mental Health: Opportunities And Challenges In Developing Intelligent Digital Therapies</em>. <a href="https://www.forbes.com/sites/bernardmarr/2023/07/06/ai-in-mental-health-opportunities-and-challenges-in-developing-intelligent-digital-therapies/">Link</a></li><li>The Lancet Digital Health, (2024). <em>Building robust, proportionate, and timely approaches to regulation and evaluation of digital mental health technologies</em>. <a href="https://doi.org/10.1016/S2589-7500(24)00215-2">Link</a></li><li>Evidence-Based Mental Health, (2022). <em>Building trust in artificial intelligence and new technologies in mental health</em>. <a href="https://mentalhealth.bmj.com/content/25/2/45">Link</a></li><li>SSRN Electronic Journal, (2024). <em>A Justifiable Investment in AI for Healthcare: Aligning Ambition with Reality</em>. <a href="https://ssrn.com/abstract=4795198">Link</a></li><li>POSTnote 737, (2025). <em>AI and Mental Healthcare: Opportunities and Delivery Considerations</em>.</li><li>Applied Sciences, (2024). <em>AI Chatbots for Mental Health: A Scoping Review of Effectiveness, Feasibility and Applications.</em> <a href="https://doi.org/10.3390/app14155889">Link</a></li><li>medRxiv, (2024). <em>AI-enabled Conversational Agent Increases Engagement with Cognitive-Behavioral Therapy</em>. <a href="https://doi.org/10.1101/2024.11.01.24316565">Link</a></li><li>Internet Interventions, (2024). <em>An Early Economic Evaluation of Kooth, a Web-Based Mental Health Platform for Children and Young People with Emerging Mental Health Needs</em>. <a href="https://doi.org/10.1016/j.invent.2024.100748">Link</a></li><li>NPJ Mental Health Research, (2024). <em>Behavioral Health and Generative AI: A Perspective on Future of Therapies and Patient Care</em>. <a href="https://www.nature.com/articles/s44184-024-00067-w">Link</a></li><li>NPJ Digital Medicine, (2022). <em>Digital biomarkers: Convergence of digital health technologies and biomarkers</em>. <a href="https://doi.org/10.1038/s41746-022-00583-z">Link</a></li><li>Nature Mental Health, (2023). <em>Economic evaluation of 27,540 patients with mood and anxiety disorders and the importance of waiting time and clinical effectiveness in mental healthcare</em>. <a href="https://doi.org/10.1038/s44220-023-00106-z">Link</a></li><li>Journal of Medicine, Surgery, and Public Health, (2024). <em>Enhancing mental health with Artificial Intelligence: Current trends and future prospects</em>. <a href="https://doi.org/10.1016/j.glmedi.2024.100099">Link</a></li><li>Lancet Digital Health, (2022). <em>From promise to practice: towards the realisation of AI-informed mental health care</em>. <a href="https://doi.org/10.1016/S2589-7500(22)00153-4">Link</a></li><li>medRxiv Preprint, (2024). <em>Combining AI and human support in mental health: a digital intervention with comparable effectiveness to human-delivered care</em>. <a href="https://doi.org/10.1101/2024.07.17.24310551">Link</a></li><li>DLA Piper Intelligence, (2025). <em>Telehealth around the world: a global guide</em>. <a href="https://www.dlapiperintelligence.com/telehealth">Link</a></li><li>World Health Organization, (2024). <em>Ethics and governance of artificial intelligence for health: Guidance on large multi-modal models</em>. <a href="https://www.who.int/publications/i/item/9789240084759">Link</a></li><li>VoiceBox, (2024). <em>Coded Companions VoiceBox Report</em>. <a href="https://voicebox.site/resources">Link</a></li><li>Cureus, (2024). <em>A Comprehensive Review on Exploring the Impact of Telemedicine on Healthcare Accessibility</em>. <a href="https://doi.org/10.7759/cureus.56789">Link</a></li><li>Nature Digital Medicine, (2024). <em>Combining AI and human support in mental health</em>. <a href="https://doi.org/10.1038/s41746-024-00987-6">Link</a></li><li>The Conversation (2024). <strong><em>1</em></strong><em> in 3 people are lonely. Will AI help, or make things worse? </em><a href="https://theconversation.com/1-in-3-people-are-lonely-will-ai-help-or-make-things-worse-217924">Link</a></li><li>WHO. (2023). <em>Artificial intelligence in mental health research: new WHO study on applications and challenges</em>. <a href="https://www.who.int/europe/news/item/06-02-2023-artificial-intelligence-in-mental-health-research--new-who-study-on-applications-and-challenges">Link</a></li><li>World Economic Forum (2024). <em>How public-private partnerships can ensure ethical, sustainable and inclusive AI development</em>. <a href="https://www.weforum.org/stories/2024/11/public-private-partnerships-ensure-ethical-sustainable-inclusive-ai-development/">Link</a></li><li>Health Foundation (2024). <em>Priorities for an AI in Health Care Strategy</em>. <a href="https://www.health.org.uk/reports-and-analysis/briefings/priorities-for-an-ai-in-health-care-strategy">Link</a></li><li>Digital Trends (2024). <em>The Best Wearables for Mental Health Tracking in 2024</em>. <a href="https://www.digitaltrends.com/mobile/best-mental-health-wearables/">Link</a></li><li>Health Innovation Network (2024). P<em>reventing Clinician Burnout: Could Ambient Voice Technology Be Key? </em><a href="https://healthinnovationnetwork.com/resources/preventing-clinician-burnout-could-ambient-voice-technology-avt-be-key/">Link</a></li><li>NHS Digital Regulations (2024). <em>Using a Chatbot to Ease Staff Shortages and Improve Patient Care</em>. <a href="https://digitalregulations.innovation.nhs.uk/case-studies/using-a-chatbot-to-ease-staff-shortages-and-improve-patient-care/">Link</a></li><li>JMIR Formative Research. (2024). <em>Effectiveness of a Mental Health Chatbot for People With Chronic Diseases: Randomized Controlled Trial</em>. <a href="https://formative.jmir.org/2024/1/e50025">Link</a></li><li>Limbic AI (2024). <em>Closing the accessibility gap to mental health treatment with a conversational AI-enabled self-referral tool</em>. <a href="https://www.limbic.ai/research/closing-accessibility-gap">Link</a></li><li>WSJ. (2024). <em>When There’s No School Counselor, There’s a Bot</em>. <a href="https://www.wsj.com/tech/ai/student-mental-health-ai-chat-bots-school-4eb1ba55">Link</a></li><li>JMIR Mental Health (2024). <em>Large Language Models for Mental Health Applications: Systematic Review</em>. <a href="https://mental.jmir.org/2024/1/e57400">Link</a></li><li>Wikipedia. (2024). <em>Artificial intelligence in mental health</em>. <a href="https://en.wikipedia.org/wiki/Artificial_intelligence_in_mental_health">Link</a></li><li>Wikipedia. (2024). <em>Digital phenotyping</em>. <a href="https://en.wikipedia.org/wiki/Digital_phenotyping">Link</a></li><li>JMIR Mental Health. (2023). <em>Conversational Agent Interventions for Mental Health Problems: Systematic Review and Meta-analysis of Randomized Controlled Trials</em>. <a href="https://www.jmir.org/2023/1/e43862">Link</a></li><li>Ada Lovelace Institute (2024). <em>Healthcare Access Denied: Ethical Concerns in Digital Health</em>. <a href="https://www.adalovelaceinstitute.org/report/healthcare-access-denied/">Link</a></li><li>Medical News Today. (2024). <em>VR therapy and its benefits for mental health</em>. <a href="https://www.medicalnewstoday.com/articles/vr-therapy">Link</a></li><li>The Guardian. (2024). <em>Virtual reality to be used in UK trial to help people beat cocaine addiction</em>. <a href="https://www.theguardian.com/science/2024/oct/18/virtual-reality-cocaine-addiction-uk-drug-overdoses">Link</a></li><li>The Guardian. (2025).<em> ‘An ideal tool’: prisons are using virtual reality to help people in solitary confinement</em>. <a href="https://www.theguardian.com/technology/2025/mar/08/vr-prison-california">Link</a></li><li>SmithBrothersUSA (2023). <em>5 Mental Health Employee Benefits Trends for 2023</em>. <a href="https://www.smithbrothersusa.com/5-mental-health-employee-benefits-trends-for-2023/">Link</a></li><li>KFF (2023). <em>Telehealth Has Played an Outsized Role in Meeting Mental Health Needs During the COVID-19 Pandemic</em>. <a href="https://www.kff.org/mental-health/issue-brief/telehealth-has-played-an-outsized-role-meeting-mental-health-needs-during-the-covid-19-pandemic">Link</a></li><li>KFF (2023). <em>5 Key Facts About Medicaid Coverage for Adults with Mental Illness.</em> <a href="https://www.kff.org/mental-health/issue-brief/5-key-facts-about-medicaid-coverage-for-adults-with-mental-illness">Link</a></li><li>TechTarget (2024). <em>Patients Indicate High Levels of Satisfaction Following Telehealth Use</em>. <a href="https://www.techtarget.com/virtualhealthcare/news/366596735/Patients-Indicate-High-Levels-of-Satisfaction-Following-Telehealth-Use">Link</a></li><li>TechTarget (2024). <em>Understanding Asynchronous Telehealth’s Benefits and Barriers to Use</em>. <a href="https://www.techtarget.com/virtualhealthcare/feature/Understanding-Asynchronous-Telehealths-Benefits-and-Barriers-to-Use">Link</a></li><li>TechTarget (2024). <em>Telehealth Patient Satisfaction On Par with In-Person Care During Pandemic</em>. <a href="https://www.techtarget.com/virtualhealthcare/news/366597341/Telehealth-Patient-Satisfaction-On-Par-with-In-Person-Care-During-Pandemic">Link</a></li><li>TechTarget (2024). <em>Understanding Key Telehealth Disparities: Patient Facing Barriers</em>. <a href="https://www.techtarget.com/virtualhealthcare/feature/Understanding-Key-Telehealth-Disparities-Patient-Facing-Barriers">Link</a></li><li>CB Insights Market Trends Report (2024). <em>AI and Mental Healthcare Opportunities</em>. <a href="https://www.cbinsights.com/research/report/digital-health-trends-2024/">Link</a></li><li>Galengrowth. (2024). <em>Global Digital Health Funding Report</em>. <a href="https://galengrowth.com/">Link</a></li><li>PR Newswire (2024). <em>Talkiatry Secures $130M Series C Funding to Mainstream Value-Based Behavioral Health Care</em>. <a href="https://www.prnewswire.com/news-releases/talkiatry-secures-130m-series-c-funding-to-mainstream-value-based-behavioral-health-care-302175461.html">Link</a></li><li>New York Post. (2025). <em>New AI tool predicts brain decline years before symptoms appear — how it could help fight dementia</em>. <a href="https://nypost.com/2025/03/08/health/new-ai-tool-predicts-brain-decline-before-symptoms-appear/">Link</a></li><li>Vator news. (2022). Mental health companies that raised funding in 2021. <a href="https://vator.tv/2022-02-02-mental-health-companies-that-raised-funding-in-2021/">Link</a></li></ol><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=3684657c4692" width="1" height="1" alt=""><hr><p><a href="https://medium.com/included-vc/why-your-next-therapist-might-be-an-ai-and-why-thats-not-a-bad-thing-3684657c4692">Why Your Next Therapist Might Be an AI — And Why That’s Not a Bad Thing</a> was originally published in <a href="https://medium.com/included-vc">Included VC</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[AI is Rewriting the Rules of Healthcare: A 2025 Investor’s Playbook]]></title>
            <link>https://medium.com/included-vc/ai-is-rewriting-the-rules-of-healthcare-a-2024-investors-playbook-649abd6a0a96?source=rss-6bd5914fe82e------2</link>
            <guid isPermaLink="false">https://medium.com/p/649abd6a0a96</guid>
            <category><![CDATA[digital-health]]></category>
            <category><![CDATA[ai-healthcare-revolution]]></category>
            <category><![CDATA[venture-capital]]></category>
            <category><![CDATA[healthcare]]></category>
            <dc:creator><![CDATA[Ugo Tomasello]]></dc:creator>
            <pubDate>Wed, 26 Feb 2025 22:18:21 GMT</pubDate>
            <atom:updated>2025-02-26T22:29:26.541Z</atom:updated>
            <content:encoded><![CDATA[<h4>The Care Crunch Chronicles</h4><p><em>Why the traditional Healthcare system can’t compete with AI — and How Investors can keep up.</em></p><figure><img alt="This image showcases AI’s role in healthcare with colorful data streams representing the flow of information into actionable insights. The top features logos of innovative startups, including Sumitomo Life, Caresyntax, Nabla, Resilience, Smart Reporting, Zipline, and Baibys. At the bottom are major tech companies like NVIDIA, IBM Watson, Microsoft, Google DeepMind, and AWS, highlighting their contributions. Included VC branding appears in the corner." src="https://cdn-images-1.medium.com/max/1024/1*7zQR7XYiqXoMiIugWlPE_g.png" /><figcaption>AI’s Impact on Healthcare: Bridging data insights with patient care through innovation by startups and tech giants</figcaption></figure><p><strong>tl;dr</strong></p><ul><li><strong>AI’s Role:</strong> Transforming healthcare with faster diagnostics, personalized treatments, and smarter operations.</li><li><strong>Case Studies:</strong> Startups like Baibys (IVF innovation) and Nabla (clinical documentation) showcase real-world impact.</li><li><strong>New Business Models:</strong> AI reshapes healthcare by creating outcome-driven solutions and automating workflows.</li><li><strong>Investment Surge:</strong> $42.5B in 2023; startups lead the charge globally.</li><li><strong>Global Impact:</strong> North America and Europe dominate; Asia and Africa innovate under unique constraints.</li><li><strong>Economic Gains:</strong> AI slashes inefficiencies, saves billions, and tackles workforce shortages.</li></ul><p><strong>. . .</strong></p><h4>Introduction</h4><p>Artificial intelligence (<strong>AI</strong>) is rapidly transforming the healthcare landscape, offering unprecedented advancements in diagnostics, treatment planning, and operational efficiency. For instance, AI-powered diagnostic tools can analyze complex medical data in seconds, significantly reducing the time required for accurate diagnosis compared to traditional methods.</p><p>A <em>study published in </em><a href="https://www.bmj.com/content/368/bmj.m689"><em>The BMJ</em></a><em> found that AI platforms provided diagnoses faster than consultants, with an average time of 2.8 minutes versus 8.5 minutes, respectively.</em></p><p>The integration of AI into healthcare systems is not only enhancing patient care but also attracting substantial investment. In 2024, digital health venture investments rebounded, driven by interest in AI, with <a href="https://www.wsj.com/articles/digital-health-venture-investment-rises-but-many-deals-have-yet-to-pay-off-d51a312f"><strong>$5.7 billion raised in the first half of the year</strong></a>. Is this surge indicating a growing recognition of AI’s potential to revolutionize healthcare delivery and generate significant returns for investors? Maybe. That’s why I am excited to talk about it.</p><figure><img alt="A bar chart showing global VC funding in healthcare AI/ML startups from 2010 to 2023, with funding rising significantly from $24 million in 2010 to a peak of $12,217 million in 2021 before declining to $4,534 million in 2023. Data source: Crunchbase." src="https://cdn-images-1.medium.com/max/1024/1*qjH5tkguPtSD6NmATM2VYA.png" /><figcaption><a href="https://www.byfounders.vc/insights/future-of-healthcare">Global Venture Capital Funding in Healthcare AI/ML Startups (2010–2023)</a></figcaption></figure><p>As we progress through 2024, we are seen the convergence of AI and healthcare as a mix of both opportunities and challenges. Investors and healthcare providers must navigate this evolving landscape to understand at first and then harness the AI’s full potential while addressing ethical, regulatory, and operational considerations.</p><p>I will explore with you how AI is restructuring the traditional healthcare system, examining the implications for investors, and providing insights from industry leaders and venture capitalists who are at the forefront of this transformation. Ready to jump into this journey?</p><h3>🔒<strong>The Healthcare system behind the wall?</strong></h3><p>AI isn’t just a buzzword in healthcare — it’s a game-changer, tackling inefficiencies and transforming patient care at every level. From turbocharged diagnostics to streamlined hospital operations, <strong>AI is rewriting the healthcare playbook</strong> in ways that were unimaginable a decade ago. But the hype and promise of a bright future come with many strings attached. The risk is to oversee the critical factors in the traditional healthcare system. This might slow down the AI integration and adoption, generating further shadows and distrust on a technology that is changing how we process reality. To shed some light, I’ve identified <strong>three key areas of innovation</strong> that need to be discussed to further understand how healthcare is moving (eventually) forward.</p><figure><img alt="This diagram, titled “AI’s Transformative Impact on Healthcare,” uses a funnel to depict AI’s role in healthcare, focusing on three areas: Smarter, Faster, Better Diagnostics (orange, clipboard icon), Healing with Data (pink, person and cross icon), and Optimizing Healthcare Operations (purple, shield icon). It illustrates how AI refines diagnostics, personalizes care, and streamlines operations." src="https://cdn-images-1.medium.com/max/1024/1*CgXvLYKc-8sAg3cE7_L90g.png" /><figcaption>AI’s Transformative Impact on Healthcare: Revolutionizing Diagnostics, Enhancing Care with Data, and Streamlining Operations for a Smarter Future</figcaption></figure><h4>1.⚕️Smarter, Faster, Better Diagnostics</h4><p>Let’s face it: traditional diagnostic methods can be slow and error-prone, often relying on manual processes that stretch resources thin. Enter AI, with its ability to process massive amounts of complex medical data in seconds. It’s not just speeding up the process; it’s elevating accuracy to new heights.</p><p>Take imaging, for example. AI-powered tools are now achieving diagnostic accuracy rates as high as 95%, outperforming many traditional methods, according to <em>The BMJ</em>. Whether it’s detecting a tumor in an MRI scan or analyzing subtle anomalies in X-rays, AI is delivering results that save lives.</p><p>But diagnostics isn’t just about images. A <a href="https://www.weforum.org/stories/2024/09/ai-diagnostics-health-outcomes">study by the World Economic Forum</a> highlighted that AI systems achieved a 93% accuracy rate in classifying heart disease, surpassing traditional diagnostic methods. GenAI is opening up new possibilities by creating synthetic patient data — essentially “fake” data with real-world applications. These datasets allow algorithms to train without running into ethical or logistical roadblocks, making it easier for healthcare providers to adopt AI safely and effectively. Tools like <a href="https://www.mdpi.com/1420-3049/29/19/4626">Generative Adversarial Networks</a> (<strong>GANs</strong>) and <a href="https://www.ibm.com/think/topics/variational-autoencoder">Variational Autoencoders</a> (<strong>VAEs</strong>) are leading this charge, synthesizing everything from medical images to electronic health records.</p><figure><img alt="A timeline detailing the evolution of AI in healthcare across four waves: Neural Networks (1990s, focusing on medical imaging), Machine Learning (2000s, identifying patterns in large datasets), Deep Learning (2010s, analyzing unstructured data), and Generative AI (2020s, leveraging natural language processing)." src="https://cdn-images-1.medium.com/max/1024/1*_UQZSSx2cI5W_9zEhEwDXQ.png" /><figcaption><a href="https://www.byfounders.vc/insights/future-of-healthcare">The Four AI Waves in Healthcare: Evolution from Neural Networks to Generative AI</a></figcaption></figure><p><a href="https://arxiv.org/abs/2407.00116">A systematic review published on <em>arXiv</em></a> highlights how these technologies are reshaping the landscape, providing a secure and scalable way to develop and validate diagnostic models. Deloitte’s research shows this innovation is already reshaping high-stakes fields like <a href="https://www2.deloitte.com/us/en/pages/about-deloitte/articles/press-releases/moffitt-cancer-center-looks-to-revolutionize-cancer-care-delivery-using-ai-and-machine-learning.html">oncology</a>, where the Cancer AI Alliance, supported by over $40 million from AWS, Deloitte, Microsoft, and NVIDIA, aims to apply responsible AI to unlock insights in cancer biology and treatment resistance and <a href="https://www2.deloitte.com/us/en/insights/industry/health-care/ai-led-transformations-in-health-care.html">emergency medicine</a>.</p><p>The <a href="https://www.precedenceresearch.com/diagnostic-testing-market">global diagnostics market is valued at an impressive $203 billion</a>. It is a vast and dynamic ecosystem where AI is quickly emerging as a “need to have” technology. Hospitals, diagnostics imaging centers, diagnostic laboratories, primary care providers and more, are taking action to gain a competitive edge in the AI-run. By 2024, the AI diagnostics segment is projected to soar to <a href="https://www.mordorintelligence.com/industry-reports/artificial-intelligence-in-diagnostics-market">$1.33 billion, driven by an eye-popping CAGR of 33%</a>, even if the real economic value generated by its adoption is hard to estimate, given that the majority of it will be calculated in “money saved” from inefficiencies by healthcare providers and national healthcare systems.</p><figure><img alt="This image is titled “AI’s Impact on Diagnostics and Treatment.” It features four dials measuring AI’s achievements: 95% accuracy in imaging diagnostics, 93% accuracy in heart disease classification, 33% growth in the AI diagnostics market, and 30.8% growth in the precision medicine market. Below, a bar chart highlights significant figures: a patient sample size of 100,000, $40M in AI investment for cancer insights, and a $203 billion global diagnostics market." src="https://cdn-images-1.medium.com/max/1024/1*wr0pL2WzhIbs1cFTYCmivw.png" /><figcaption>AI boosts diagnostics: 95% imaging accuracy, 93% heart disease detection, and $203B market growth</figcaption></figure><h4>2. 💉Healing with Data📈: Redefining Treatment, One Patient at a Time</h4><p>AI is not just diagnosing conditions with increasing accuracy, it’s also influencing how treatments are planned, delivered, and optimized to meet the unique patient’s needs. By analyzing vast datasets encompassing genetic profiles, patient histories, environmental factors, and real-time health metrics, <strong>AI empowers clinicians to create personalized treatment</strong> pathways that improve efficacy and reduce risks.</p><p>One of AI’s most striking contributions is in <strong>precision medicine</strong>. For instance, AI algorithms can predict patient responses to specific medications, allowing physicians to bypass costly trial-and-error approaches, by analyzing patient’s genetic and clinical profile. This targeted methodology could save the healthcare industry billions annually by reducing ineffective treatments and complications. The precision medicine market, reflecting this trend, is expected to raise from <a href="https://www.gminsights.com/industry-analysis/artificial-intelligence-in-precision-medicine-market">$1.2 billion in 2022 to $17.1 billion by 2032</a>, growing at an impressive CAGR of 30.8%.</p><figure><img alt="This infographic highlights the role of AI in the precision medicine market. It shows a 30.8% compound annual growth rate (CAGR) from 2023 to 2032, with the market valued at $1.2 billion in 2022 and projected to reach $17.1 billion by 2032. Key segments include deep learning (36.5% share), software (53.4% share), and oncology ($5.7 billion by 2032). North America leads with 51.9% market share in 2022. Data source: Global Market Insights." src="https://cdn-images-1.medium.com/max/627/1*G6MD1mGS-35fg37YxVwLZg.png" /><figcaption><a href="https://www.gminsights.com/industry-analysis/artificial-intelligence-in-precision-medicine-market">AI in Precision Medicine: $1.2B market (2022), 30.8% CAGR, $17.1B forecast (2032), led by North America (51.9% share)</a></figcaption></figure><p>Beyond precision, <strong>AI is enhancing the efficiency of treatment</strong> <strong>delivery.</strong> AI-powered virtual assistants now aid in medication management, patient follow-ups, and even surgical planning. Hospitals leveraging these technologies in <a href="https://www.advisory.com/daily-briefing/2023/03/15/medical-errors"><strong>medication management</strong></a>, by scanning electronic medical records to identify potential errors, improved patient safety but also optimizes resource allocation. In <strong>critical care</strong>, <a href="https://www.advisory.com/daily-briefing/2022/04/12/ai-hospitals">a study with 160 hospitals</a> showed that AI-enabled tools to continuously monitor patient data, can alert clinicians to potential issues like sepsis hours before symptoms appear —<strong> a capability shown to reduce mortality rates by up to 30%</strong>. If you want to discover more about treatment delivery, you might want to check the Owen Keating’s article: <a href="https://medium.com/included-vc/virtual-hospitals-real-returns-a-venture-into-virtual-hospital-wards-b566f2cc1401"><strong>Virtual Hospitals, Real Returns: A Venture into Virtual Hospital Wards</strong></a><strong>.</strong></p><p><a href="https://www.mdpi.com/1999-4923/16/10/1328"><strong>AI’s impact extends also into drug discovery</strong></a>, where machine learning algorithms predict molecular interactions and optimize drug formulations. These advancements accelerate the drug development process, reduce costs, and pave the way for therapies that are finely tuned to individual patient needs. If you want to further dive into the impact of AI-driven Drug Discovery and Development, I suggest you to read my previous Medium article that analyzes it in detail: <a href="https://medium.com/included-vc/investing-in-biotechs-ai-gold-rush-what-smart-money-needs-to-know-about-the-future-of-medicine-05d2d9841ee1"><strong>Biotech’s AI Gold Rush: What Smart Money needs to learn to invest on the Future of Medicine.</strong></a></p><figure><img alt="Illustration of a human head profile composed of geometric, digital patterns, symbolizing AI and biotechnology integration. Two digital, pill-like shapes on the left emit colorful triangular particles, suggesting the concept of drug discovery through AI. The image is surrounded by logos of notable biotech and venture capital companies, including a16z, Kurma Partners, V-Bio Ventures, Forbion, VIB, Insilico Medicine, and insitro, representing key players in the biotech and AI investment landscape" src="https://cdn-images-1.medium.com/max/1024/1*HkrP95n0JAlZOiU3rIqgVw.png" /><figcaption><a href="https://medium.com/included-vc/investing-in-biotechs-ai-gold-rush-what-smart-money-needs-to-know-about-the-future-of-medicine-05d2d9841ee1">The Convergence of AI and Life Sciences: Top Investors and Innovators Shaping the Future of Drug Discovery and Biotechnology</a></figcaption></figure><h4>3. 🏥Optimizing Healthcare Operations with AI: Efficiency Redefined🛠️</h4><p>Healthcare operations is one of the biggest economic burden of the healthcare system. From patient scheduling and claims processing to electronic health record (EHR) management, these tasks are widely inefficient and generate a substantial economic loss. Automated systems, namely Robotic Process Automation (RPA) are projected to save the global healthcare industry a staggering $200 to $360 billion over the next five years, <a href="https://www.marketsandmarkets.com/Market-Reports/artificial-intelligence-healthcare-market-54679303.html">constituting 5% to 10% of healthcare spending.</a></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*wiDlzA6kPKVPKCMQdbt2dg.png" /><figcaption><a href="https://automationedge.com/home-health-care-automation/blogs/ai-and-rpa-in-healthcare/">Top AI &amp; RPA use cases</a></figcaption></figure><p>AI is also tapping on the supply chain management. Predictive analytics powered by machine learning have <a href="https://www.linkedin.com/pulse/30-less-overstock-stockouts-how-predictive-analytics-babenko-ph-d--sigge/"><strong>reduced hospital stockouts by 20–30% while cutting inventory costs by as much as 15%</strong></a>. Staffing optimization, a pressing challenge given the projected <a href="https://www.who.int/health-topics/health-workforce#tab=tab_1">global shortfall of 10 million healthcare workers by 2030</a>, is another area of impact. Predictive algorithms analyze patient inflow trends, enabling facilities to allocate staff effectively, minimize overtime, and reduce burnout.</p><p>Operational efficiency extends even further to hospital workflows. AI-powered task prioritization systems have <a href="https://www.linkedin.com/pulse/leveraging-ai-reduce-hospital-wait-times-andre-ripla-pgcert-qhg9e/">cut the average wait time for critical interventions, visit or surgery by 20–30%, according to recent studies in several hospitals (Mayo Clinic, John Hopkins Hospital and more).</a></p><p>The combined effect of these AI-powered innovations is radically transformative, changing the healthcare system from inside.</p><blockquote><a href="https://cepr.org/voxeu/columns/what-happens-when-ai-comes-healthcare">Robotic Process Automation (RPA) are projected to save the global healthcare industry a staggering $200 to $360 billion annually.</a></blockquote><figure><img alt="This bar chart shows the global healthcare automation market growing from $49.5B in 2023 to $119.5B by 2033, with a 9.2% CAGR. Key segments include therapeutic automation, lab &amp; pharmacy automation, medical logistics &amp; training, and diagnostics &amp; monitoring, each shown in distinct colors. The chart highlights steady growth in all segments, with 2033 dominated by diagnostics &amp; monitoring. The data is sourced from market.us, emphasizing automation’s expanding role in healthcare." src="https://cdn-images-1.medium.com/max/1024/1*u5356VdAYnhzqHth2XIb7g.png" /><figcaption><a href="https://market.us/report/healthcare-automation-market/">Global Healthcare Automation Market</a></figcaption></figure><p>But the benefits go beyond cost savings — these technologies free up providers to focus on patient care, improve accessibility, and raise the overall quality of healthcare delivery. AI is not just making operations smarter; it’s creating a more sustainable and patient-focused future for global healthcare systems.</p><h4>BAIBYS Case Study: The first “automatic gear” in the world of IVF</h4><figure><img alt="The Baibys™ logo features the brand name in a mix of blue and gray text. The “B” incorporates a stylized silhouette of a person in profile. Below the name, the tagline “Select with care” is written in blue. This logo conveys a focus on precision and thoughtful decision-making in healthcare." src="https://cdn-images-1.medium.com/max/867/1*hUfCRh7lSDYArkuPMjWs-A.png" /><figcaption><a href="https://www.baibys.com/about-us/">Baibys</a></figcaption></figure><p>In the quest to understand how AI is transforming healthcare, I spoke with <a href="https://www.linkedin.com/in/yaronsilberman/"><strong>Yaron Silberman</strong></a>, CEO of <a href="https://www.baibys.com/"><strong>Baibys</strong></a>, a startup pioneering AI-driven sperm selection for in-vitro fertilization (IVF). Baibys is a fantastic example of AI used a multiple levels integrating software and robotics, to streamline the medical procedure workflow, improving clinical positive output and democratizing a costly medical option. In this first part I will explore with Yaron how they leveraged AI to solve a demographic global challenge.</p><p><strong><em>What prompted Baibys to focus on improving sperm selection in IVF procedures?</em></strong></p><p><strong>Yaron</strong>: “We identified a significant challenge with traditional sperm selection methods. The criteria for what constitutes a normal sperm cell are extremely strict — the WHO manual details this extensively. Sperm cells are tiny, and they’re in constant motion. Assessing them manually requires embryologists to make complex calculations in real-time, which is nearly impossible with complete accuracy.”</p><p><strong><em>How does your AI technology address this challenge?</em></strong></p><p><strong>Yaron</strong>: “Our solution utilizes a <strong>multi-layered AI algorithm</strong>. The first layer detects sperm cells in microscopic images, filtering out artifacts that can resemble sperm cells. The second layer keeps the sperm in focus despite their movement along the Z-axis. The <strong>most critical layer</strong> is our deep learning model — specifically, we use VGG-16 — that <strong>evaluates whether a sperm cell meets the strict morphological criteria set by the WHO</strong>.</p><p>“We trained our AI on a dataset of 15,000 images labeled by at least two embryologists to ensure consensus, given that professionals only agree on sperm quality 40–70% of the time. This approach enhances the reliability of our assessments compared to manual evaluations.”</p><p><strong><em>What advantages does this bring to IVF clinics and patients?</em></strong></p><p><strong>Yaron</strong>: “By automating sperm selection, we’re alleviating one of the most tedious tasks embryologists face. <strong>This not only reduces the potential for human error but also frees up skilled professionals to focus on more critical aspects of patient care</strong>. Clinics can significantly increase their throughput — handling more cases without the need for additional staff or resources.</p><p>“Considering that the demand for IVF is <strong>four to five times higher</strong> than the available services worldwide, our technology helps bridge that gap. It enhances workflow efficiency, which can lead to reduced costs and increased accessibility for patients seeking fertility treatments.”</p><p><strong><em>How have embryologists and clinics responded to your technology?</em></strong></p><p><strong>Yaron</strong>: “The feedback has been overwhelmingly positive. When we first demonstrated our AI’s capabilities, embryologists were astounded by how quickly and accurately it could perform tasks that are incredibly challenging manually. They appreciate that the system handles a part of the process they often find frustrating.</p><p>“Moreover, since our system integrates seamlessly into existing workflows without requiring extra real estate, extensive training, or maintenance, clinics find it easy to adopt. <strong>It’s a cost-effective solution that doesn’t disrupt their current operations</strong>.”</p><p><strong><em>Looking ahead, how do you envision the impact of Baibys on reproductive medicine?</em></strong></p><p><strong>Yaron</strong>: “We believe that AI-driven solutions like ours will set new standards in IVF treatments. Clinics that adopt such technologies will not only improve success rates but also enhance patient satisfaction by offering more efficient services. In the long term, as patients become more informed, they may prefer clinics utilizing advanced technologies, making it essential for providers to adapt.”</p><blockquote>“<strong>Ultimately, our goal is to contribute to higher success rates and healthier outcomes, reducing the emotional and financial strain on families undergoing fertility treatments.</strong>”</blockquote><h3><strong>Market Trends: AI’s Meteoric Rise in Healthcare</strong> 🔍📊</h3><p>The healthcare sector is undergoing a fundamental transformation, catalyzed by advancements in AI technologies. This revolution transcends mere operational efficiency; it represents a reconfiguration of care delivery, decision-making processes, and resource management. We have just discussed three key innovation areas, and how they are reshaping operations and workflow in the healthcare sector. Now it’s time to deep dive into some market trends and analytics to understand the weight of the AI revolution.</p><p>Starting with the big picture, the global AI healthcare market, valued at <a href="https://www.fortunebusinessinsights.com/industry-reports/artificial-intelligence-in-healthcare-market-100534">$19.5 billion in 2023, is forecast to reach <strong>$27.7 billion in 2024</strong> and $490.9 billion by 2032, reflecting a staggering CAGR of 43.2%</a>.</p><p><strong>The Key Catalysts of Growth:</strong></p><ul><li>The escalating <strong>prevalence of chronic illnesses</strong> necessitating advanced diagnostic and therapeutic approaches.</li><li>Heightened demand for <strong>personalized medicine</strong> driven by genomic insights and patient-specific interventions.</li><li>A growing <strong>shortages of healthcare professionals</strong> worldwide, necessitating efficiency-enhancing innovations.</li></ul><p>But, how this catalyst translate into new market trends?</p><figure><img alt="A fishbone diagram titled “AI’s Meteoric Rise in Healthcare” highlights three factors: AI Value Chain (back-end data and front-end care), Economic Drivers (cost-saving and startup funding), and Technology Adoption (emerging regions and payor-provider sync)." src="https://cdn-images-1.medium.com/max/1024/1*RObR0aC9M86mk3BlQ5Ad2A.png" /><figcaption>AI’s Meteoric Rise in Healthcare: Key drivers include the AI value chain, economic factors, and tech adoption</figcaption></figure><h4>1. 🚀 The AI Value Chain: Connecting Back-End Intelligence to Front-End Care</h4><p>AI is revolutionizing healthcare, <strong>end-to-end</strong>. From crunching massive datasets at lightning speed to delivering life-saving diagnoses and personalized treatment plans, AI creates a seamless <strong>pipeline between insights and action</strong>.</p><ul><li><strong>Back-End Brilliance</strong>: The heavy lifting happens here. AI-powered tools analyze oceans of data — X-rays, genomic sequences, and hospital workflows — and translate them into actionable insights that healthcare professionals rely on.</li></ul><p>In 2024, AI radiology platforms, which represent on the biggest AI-adopting segment, cut <a href="https://healthimaging.com/topics/artificial-intelligence/commercially-available-ai-reduces-radiologists-workload-34-certain-screening-settings"><strong>breast cancer screening, reducing radiologist workload by 34%</strong></a>. This alone saves both <strong>lives</strong> and <strong>resources</strong>. According to <a href="https://www3.weforum.org/docs/WEF_Transforming_Healthcare_2024.pdf"><strong>WEF’s 2024 Report</strong></a>, back-end AI applications in diagnostics and imaging remain the largest AI healthcare market segment, with adoption accelerating in both developed and emerging markets.</p><ul><li><strong>Front-End Impact</strong>: AI tools like <strong>virtual health assistants</strong>, wearables, and personalized dashboards are turning patients into <strong>proactive participants</strong> in their own care. Real-time monitoring of chronic conditions — like diabetes or heart disease — means intervention happens before complications arise.</li></ul><p>AI-integrated glucose monitors, such as <a href="https://www.fiercebiotech.com/medtech/roche-introduces-ai-powered-diabetes-tracker-predict-blood-sugar-highs-and-lows"><strong>Roche’s Accu-Chek SmartGuide</strong></a><strong>,</strong> provide continuous glucose monitor (CGM) AI-based predictions of blood glucose levels from 30 minutes up to two hours in advance, assisting users in managing potential hypoglycemic events.</p><figure><img alt="This diagram highlights the interplay between front-end tools (patient communication, virtual health, monitoring, and chronic disease management) and back-end systems (data processing, infrastructure, diagnostics, and supply chain management) in healthcare. It also maps the patient journey across “Before the Walls,” “In the Walls,” and “After the Walls,” showcasing AI’s role in integrating care. Icons and colors differentiate each category, with Included VC branding at the bottom." src="https://cdn-images-1.medium.com/max/1024/1*0_dfLqvGAv_667YSluLfUw.png" /><figcaption>AI in Healthcare: Connecting front-end patient tools with back-end systems across the care journey</figcaption></figure><h4>2. 🌍 Convergence of Technology and Adoption</h4><p>AI is no longer futuristic — it’s happening now. <strong>But adoption isn’t uniform</strong>. Some regions and sectors are leading the charge, while others are catching up.</p><p><strong>Providers are Paving the Way</strong>: By 2025, it’s anticipated that <a href="https://www.dialoghealth.com/post/ai-healthcare-statistics"><strong>90% of hospitals will utilize AI-powered technology</strong></a> for early diagnosis and remote patient monitoring, primarily in <strong>radiology</strong> and <strong>oncology</strong>. Why? These areas see the <strong>clearest ROI</strong> in both patient outcomes and cost savings.</p><p><strong>Payors Are Joining the Movement: </strong>Payors (insurance companies and health systems) are increasingly adopting AI to <strong>streamline claims processing</strong>, reduce fraud, and identify at-risk populations. AI-powered platforms process claims faster and more accurately, cutting costs associated with manual reviews. US <strong>Anthem &amp; UnitedHealth Group </strong><a href="https://julienflorkin.com/business/insurance/ai-fraud-detection">reduced fraudulent claims by 30% and 40% respectively</a>, saving millions annually.</p><blockquote><a href="https://research.aimultiple.com/ai-health-insurance">Healthcare fraud costs about $300 billion annually in the U.S</a></blockquote><p><strong>Population Health Management</strong>: AI helps payors predict chronic disease risks, enabling proactive interventions. This reduces long-term costs and aligns with value-based care models. <a href="https://asia.insuretechconnect.com/articles/transforming-insurance-ai-sumitomo-lifes-journey-toward-personalized-preventive-care">In Japan, for example, <strong>insurers</strong>, like Sumitomo Life, <strong>are integrating AI into wearable devices to track and incentivize healthy behavior</strong>, reducing claims for preventable conditions.</a></p><p>The pace of AI adoption varies widely also <strong>across geographical regions</strong>, influenced by local infrastructure, government policies, and economic priorities.</p><p><strong>Regions on the Rise</strong>: <strong>North</strong> <strong>America</strong> and <strong>Europe</strong> regions lead the global AI race due to robust healthcare infrastructure, established AI companies, and regulatory clarity, with a <a href="https://www.allaboutai.com/resources/ai-statistics/healthcare/#clinical-and-diagnostic-improvements-resulting-from-ai-in-healthcare"><strong>cumulative AI-healthcare market share of +80%</strong></a>.</p><p><strong>Asia, the future powerhouse:</strong> Countries like <strong>India</strong>, <strong>China</strong>, and <strong>Singapore</strong> are making massive investments in digital health, with <a href="https://documents1.worldbank.org/curated/en/099081723223028498/pdf/P1750750936a760320bfb9066084a6084bf.pdf">India’s <em>National Digital Health Mission</em></a> aiming to connect every citizen with AI-driven healthcare services by 2030</p><p><strong>Innovating under constraints:</strong> <strong>Africa</strong> is leveraging AI to tackle healthcare challenges like limited infrastructure and access. In <strong>Kenya</strong>, platforms like <a href="https://www.voanews.com/a/ai-enhances-maternal-health-care-in-kenya-experts-say/7789467.html"><em>Mama’s Hub</em></a> improve maternal outcomes by providing prenatal care recommendations, while <strong>Rwanda</strong> uses <a href="https://news.mit.edu/2022/protecting-maternal-health-rwanda-0918">AI to identify high-risk pregnancies</a>. Tools in <strong>Ghana</strong> and <strong>Nigeria</strong> aid in early disease diagnosis, and <strong>AI-driven drones</strong> like <a href="https://www.africanews.com/2024/10/25/zipline-partners-nigeria-for-improved-healthcare-delivery"><em>Zipline</em></a> deliver medical supplies to remote areas, cutting delivery times by 80%. With over 650 million mobile users, apps like <em>Ada</em> and <em>Babylon Health</em> offer affordable <strong>virtual consultations</strong>. Africa’s healthcare <a href="https://www.statista.com/outlook/tmo/artificial-intelligence/africa">AI market is projected to grow over 25% annually through 2030</a>, showcasing its potential to transform healthcare access.</p><figure><img alt="This infographic shows AI in Healthcare Market Share by Region and AI Adoption by Healthcare Sector. North America leads with 60%, followed by Europe (20%), Asia-Pacific (15%), and RoW (5%). AI adoption is highest among Healthcare Providers (45%), followed by Pharma &amp; Life Sciences (35%) and Healthcare Payors (20%). The visual uses balloons for regions and bar charts for sector adoption. Credit: Ugo Tomasello, Included VC." src="https://cdn-images-1.medium.com/max/1024/1*9ZgC_Z-vsQyk86Mw8UezZw.png" /><figcaption>AI Healthcare market shares by geographic clusters and end users</figcaption></figure><h4><strong>3. </strong>💰 <strong>The Economic Engine: </strong>Cure Providers before the patients.</h4><p>The key word in healthcare is to save money. Healthcare is one of the most expensive and inefficient sector for each government. With a growing old population and raise of chronic and degenerative diseases, there are zero chances to imagine a functional healthcare system without foundational reshape. We could call it the <strong>Care Crunch</strong>. One of the most scaring fact is that globally, the healthcare spend is rising faster than the GDP (which represent the economic power of a country).</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*TFeolqSdsqg1XTh4zefcdw.png" /><figcaption><a href="http://www3.weforum.org/docs/WEF_Transforming_Healthcare_2024.pdf">Healthcare spend vs GDP and Worldwide healthcare workers shortage</a></figcaption></figure><p><strong>Cut Costs, Save Lives: </strong>otherwise said, “Save Lives, Cut Costs. While we already seen some key areas of innovation, let’s settle now on six key ares of economic loss and how they are affecting the healthcare system. Let’s break down some numbers.</p><ul><li>The<strong> Administrative Complexity </strong>including billing and insurance-related activities, is very inefficient and account for <a href="https://jamanetwork.com/journals/jama/fullarticle/2785479"><strong>15% to 25% of total national health care expenditures</strong> for $600 billion to $1 trillion per year</a>.</li><li>Many<strong> Hospital Readmissions are avoidable </strong>resulting from inadequate discharge planning or follow-up care. In 2017, potentially preventable adult hospital stays resulted in $33.7 billion in aggregate hospital costs.</li><li><strong>Misdiagnosis and Medical Errors,</strong> leading to unnecessary treatments, increased healthcare costs and unnecessary deaths. <a href="https://pubmed.ncbi.nlm.nih.gov/29763131/">A study revealed that financial harm of these errors costly roughly <strong>$20 billion/year</strong></a>.</li><li>Despite<strong> </strong>some positive examples,<strong> Fraud, Waste, and Abuse still represent a huge burden. </strong>Each year, healthcare payers and providers in the US spend about <a href="https://www.americanprogress.org/article/excess-administrative-costs-burden-u-s-health-care-system/"><strong>$496 billion on billing</strong> and insurance-related (BIR) costs</a>.</li><li><strong>Poor Chronic Disease Management </strong>leads to increased hospitalizations and healthcare utilization. While it is hard to estimate the cost of the poor management, the whole chronic disease management take <a href="https://www.cdc.gov/chronic-disease/data-research/facts-stats/index.html">90% of the total healthcare expediture in US, which amounts to $4 trillions.</a></li><li><strong>Outdated Supply Chain Management</strong> results in wasted resources and increased operational costs. <a href="https://www.supplychainbrain.com/articles/39143-how-to-cut-spending-waste-from-healthcare-supply-chains"><strong>About 25% can be considered waste</strong>, accounting for $25 billions/year</a>.</li></ul><figure><img alt="A visual representation showing five circular data points highlighting healthcare costs, including $600 billion (20% of healthcare expenditure), $34 billion, $20 billion annually, $496 billion (90% of healthcare expenditure), and $25 billion (25% of wasted resources)." src="https://cdn-images-1.medium.com/max/1024/1*5j-d0xnc84SkAyiAiXEEaw.png" /><figcaption>Key statistics on healthcare expenditures and inefficiencies.</figcaption></figure><p><strong>The Startup Surge and AI Pioneers in Healthcare: </strong>In 2023, AI startups secured <strong>$42.5 billion in funding</strong>, with generative AI companies accounting for <strong>48% of the total</strong>. This surge highlights a significant <strong>shift in investment focus</strong> from clinical tools to operational AI, as genAI tools have proven effective in <strong>non-clinical operations</strong> such as clinical documentation, workflow automation, and resource management.</p><p><strong>Few emerging AI-driven healthcare stars are leading the investment deals in Europe.</strong></p><p>Berlin-based <a href="https://caresyntax.com/"><strong>Caresyntax</strong></a> leverages AI to improve surgical outcomes through real-time data-driven decisions, helping surgeons to reduce complications and improve operational efficiency. They raised <a href="https://techfundingnews.com/ai-for-healthcare-top-funding-rounds-of-europe-in-2024/"><strong>$299 millions</strong> from a big investor community, including Symbiotic Capital, MTIP AG, BIONIQ Capital, PFM Health Sciences, Pictet Alternative Advisors, surgical.ai, BlackRock Innovation Capital, Aescuvest, Optum Ventures, Cure Capital, Relyens Group, Vesalius Biocapital, Lauxera Capital, Plug &amp; Play Growth Fund, and ProAssurance Corp</a>.</p><p><a href="https://www.smart-reporting.com/"><strong>Smart Reporting</strong></a>, a German startup, secured <a href="https://techfundingnews.com/ai-for-healthcare-top-funding-rounds-of-europe-in-2024/"><strong>$54.8 millions </strong>(from Bayern Kapital and TVM Capital Life Science)</a> to advance its AI-powered radiology reporting tools, improving diagnostic workflows and data accuracy, combining structured reporting templates with NLP.</p><p><a href="https://www.resilience.care/"><strong>Resilience</strong></a>, a France-based company, focuses on AI-driven solutions for cancer patient management. By combining patient-reported outcomes (PROs) with clinical data, Resilience optimizes <strong>care pathways</strong> for oncology patients, offering real-time insights into treatment efficacy and symptom management. They have secured <a href="https://techfundingnews.com/ai-for-healthcare-top-funding-rounds-of-europe-in-2024/"><strong>$74.5 millions</strong> from Seaya, Singular, Cathay Innovation, Picus Capital, and Red River West</a> and others.</p><p>Finally, <strong>Nabla</strong>, a Paris-based startup, is pioneering the field with a focus on <strong>clinical documentation</strong>. Nabla’s platform automates the creation of clinical notes reducing administrative burdens for physicians and integrating them into EHR. The company raised <a href="https://techfundingnews.com/ai-for-healthcare-top-funding-rounds-of-europe-in-2024/"><strong>€42.7 millions</strong> from Cathay Innovation and ZEBOX Ventures</a> and more, to scale its innovative solution across Europe.</p><p>In this fast-pacing scenario, big tech players are not just hanging around. <strong>Microsoft (</strong>Azure OpenAI Service<strong>) </strong>and <strong>Amazon (</strong>AWS<strong>) </strong>are accelerating AI adoption by integrating generative AI tools into <strong>EHRs</strong> and offering AI-powered clinical documentation services. Healthcare providers, especially smaller clinics, are accessing cutting-edge technology through <strong>AI-as-a-Service platforms</strong>, which offer subscription-based models that eliminate the need for substantial upfront investments.</p><p>Other giant tech which are strongly contributing to the AI integration are <strong>IBM Watson Health</strong>, <strong>Google DeepMind</strong>, and <strong>Philips. </strong>IBM Watson Health supports clinical decision-making and optimizes patient data analysis, Google DeepMind advances AI applications for diagnostics and predictive care, and Philips integrates AI into imaging systems and connected care solutions, delivering innovation across healthcare ecosystems.</p><figure><img alt="A comprehensive market map showcasing healthcare AI startups categorized by areas such as medical coding, call center operations, medical imaging, clinical search engines, decision support, patient consultations, clinical datasets, data structuring, patient monitoring, and consumer-facing AI." src="https://cdn-images-1.medium.com/max/1024/1*91xmJuDQdidunBZKgjks4w.png" /><figcaption><a href="https://www.byfounders.vc/insights/future-of-healthcare">Healthcare AI Market Map: Categorization of AI-driven startups across clinical and non-clinical applications</a></figcaption></figure><blockquote>There are zero chances to imagine a functional healthcare system without foundational reshape. We could call it the <strong>Care Crunch.</strong></blockquote><h4>Investor Spotlight: Heal Capital’s Approach to AI in Healthcare</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/528/1*wB_wevo4IzcSad6IlEE9Lw.png" /><figcaption><a href="https://www.healcapital.com/">Heal Capital logo</a></figcaption></figure><p>To better understand the investment landscape for AI-driven healthcare startups, I spoke with <a href="https://www.linkedin.com/in/vahid-asimi/"><strong>Dr.</strong> <strong>Vahid Asimi</strong></a>, an early-stage investor at <a href="https://www.healcapital.com/"><strong>Heal Capital</strong></a>, a HealthTech-focused VC fund investing across Europe. Vahid offered valuable perspectives on what makes AI startups attractive to investors and how Heal Capital assesses the real-world impact of AI applications.</p><p><strong>What attributes do you look for in healthcare startups using AI?</strong><br>“At Heal Capital, we focus on startups that operate at the <strong>application layer</strong> of AI. This means companies that use AI to deliver tangible, real-world outcomes — whether it’s improving diagnostics, streamlining patient flows, or automating care delivery. We don’t go into the infrastructure or foundation model layers. If the solution impacts millions of patients or transforms healthcare operations, that raises our attention.”</p><p><strong>Can you share an example of a companies that reflect this focus?</strong><br>“One example is our portfolio company <strong>Informedica</strong>. They’re unique because they started building their AI-based diagnostics platform before the generative AI craze. Their approach to AI has always been pragmatic and method-driven, and now, with advancements in large language models, they’re well-positioned to scale further.</p><p>“We’re also seeing growing interest in <strong>federated learning</strong> platforms like <strong>Fears</strong>, which connect <strong>Pharma companies and healthcare providers in a decentralized way</strong>. They enable data training while preserving patient privacy — a huge challenge in AI adoption for healthcare.”</p><p><strong>Which trends are shaping the adoption of AI in healthcare?</strong><br>“Right now, if a company has any sort of <strong>proprietary dataset</strong>, they’re finding ways to mine that data using AI. That’s a significant trend. We’re also seeing the rise of <strong>TechBio</strong>, where AI is used to generate scientific hypotheses and automate lab experiments. This AI-driven productivity allows researchers to produce a higher volume of ideas and data, accelerating R&amp;D cycles.</p><p>“In healthcare software, AI is also unlocking <strong>new</strong> <strong>business models</strong>. For example, an AI-powered platform doesn’t just provide tools; it can replace parts of the manual workforce. That’s a step change in value creation — you’re no longer just selling software, but outcomes.”</p><p><strong>How do you balance the hype of AI with realistic investment returns?</strong><br>“It’s easy to get swept up in the hype, but at Heal Capital, we focus on what problem the startup is solving. Is the AI tool actually creating value? If the answer is yes — whether through improved efficiency, better diagnostics, or cost savings — then we dig deeper.</p><p>“We don’t assume AI alone adds value. The real question is: <strong>What impact does it deliver?</strong> That’s what matters.”</p><h3>🦋 Conclusions: AI is Healthcare’s Catalyst for Change</h3><p>AI is no longer a vision for the future — it is here, revolutionizing how healthcare operates. From precision diagnostics to personalized treatments and operational efficiency, AI is addressing critical challenges that have long hindered healthcare systems. But<strong> its true potential lies beyond automation</strong>: AI is reshaping the way healthcare providers deliver value and <strong>how patients experience care</strong>.</p><p>For <strong>providers</strong>, AI enables smarter workflows, reducing inefficiencies and improving patient outcomes. Hospitals leveraging AI for diagnostics and predictive analytics are <strong>not just saving costs but also saving lives</strong>. By adopting tools like robotic process automation and AI-powered assistants, healthcare organizations can mitigate workforce shortages and enhance care delivery.</p><p>For <strong>startups</strong>, AI is unlocking entirely new business models. Tools like those developed by <strong>Baibys</strong> and <strong>Nabla</strong> demonstrate how AI can automate complex processes and create scalable solutions. These innovations enable providers to move from fee-for-service models to outcome-based care, aligning with the global push for value-driven healthcare.</p><p>For <strong>investors</strong>, the opportunity is immense but requires sharp discernment. With <strong>$42.5 billion raised in 2023 alone</strong>, AI in healthcare is one of the most lucrative sectors. Successful investments will focus on startups solving real-world problems, such as improving diagnostics, streamlining operations, or creating personalized treatments. Companies that leverage proprietary datasets and integrate AI into existing healthcare workflows stand to deliver the most value.</p><p>Globally, the <strong>healthcare AI adoption curve</strong> is steep. North America and Europe lead the way, but emerging markets in Asia and Africa show how AI can innovate even under resource constraints. This geographic diversity underscores the transformative potential of AI across all corners of the globe.</p><p>Ultimately, <strong>AI is more than a technology — it is a strategic imperative.</strong> Providers must adopt it to stay relevant, startups must innovate to stay competitive, and investors must navigate the hype to capture real value. The convergence of AI and healthcare is not just about improving efficiency; it is about reimagining the future of care. For those ready to embrace it, the rewards are transformative.</p><p><strong>. . .</strong></p><h3>🧩 About me</h3><p>With a background spanning <strong>R&amp;D in Molecular Neurobiology</strong>, Clinical Development Management, I am earning my stripes in Venture Capital, because I love to explore new perspectives to evaluate Tech innovation and <strong>investments in Life Sciences</strong>. My commitment to transforming scientific insights into tangible health-tech solutions fuels my passion for this evolving AI-healthcare intersection.</p><p>If you want to discuss these exciting trends further — please feel free to reach out on <a href="https://www.linkedin.com/in/ugo-tomasello/">LinkedIn</a> or via <a href="http://ugo.tomasello@gmail.com/">email</a>.</p><p>👉 <strong>You can find me on </strong><a href="https://www.linkedin.com/in/ugo-tomasello/"><strong>LinkedIn</strong></a><strong> and keep up to date with my latest insights on </strong><a href="https://medium.com/@ugotomasello"><strong>Medium</strong></a></p><p>👉🏼 <strong><em>Follow </em></strong><a href="https://included.vc/"><strong><em>Included VC</em></strong></a><strong><em> on </em></strong><a href="https://twitter.com/included_vc"><strong><em>Twitter</em></strong></a><strong><em> &amp; </em></strong><a href="https://www.linkedin.com/company/includedvc/"><strong><em>LinkedIn</em></strong></a></p><p><strong>#includedvcmafia</strong></p><p><strong>. . .</strong></p><p>🙏 <strong>Special thanks</strong></p><p>A special thank you note to everyone who contributed to the development of this article (the Included VC team, Included VC Delegates, and all the VCs whose work and testimonies allowed me to develop extensive research on the topic). Namely, I would like to thank <a href="https://www.linkedin.com/in/yaronsilberman/">Dr. Yaron Silberman</a>, <a href="https://www.linkedin.com/in/vahid-asimi/">Dr. Vahdi Asimi</a> and <a href="https://www.linkedin.com/in/nikitathakrar/">Nikita Thakrar</a>.</p><p><strong>References</strong></p><ol><li>Medium. (2024). <em>2024 Health Trends: 4 experts discuss key innovation priorities</em>. <a href="https://medium.com/redesign-health-blog/2024-health-trends-4-experts-discuss-key-innovation-priorities-681a3d593167">Link</a></li><li>Sidley. (2024). <em>Top trends in the European digital health AI market</em>. <a href="https://datamatters.sidley.com/2024/10/09/top-trends-in-the-european-digital-health-ai-market/">Link</a></li><li>Daimagister. (2024). <em>2024 Trends in Digital Health: A Deep Dive</em>. <a href="https://www.daimagister.com/resources/health/">Link</a></li><li>Sequoia Capital. (2023). <em>Generative AI for healthcare: A transformative perspective</em>. <a href="https://www.sequoiacap.com/article/generative-ai-for-healthcare-perspective/">Link</a></li><li>Deloitte. (2024). <em>The Generative AI Dossier</em>. <a href="https://www2.deloitte.com/content/dam/Deloitte/us/Documents/consulting/us-ai-institute-gen-ai-use-cases.pdf">Link</a></li><li>Deloitte. (2024). <em>Transforming healthcare with artificial intelligence</em>. <a href="https://www.deloitte.com/content/dam/assets-shared/docs/industries/life-sciences-health-care/2024/global-transforming-health-care-with-artificial-intelligence.pdf">Link</a></li><li>Fortune Business Insights. (n.d.). <em>Artificial intelligence in healthcare market size and forecast</em>. <a href="https://www.fortunebusinessinsights.com/industry-reports/artificial-intelligence-in-healthcare-market-100534">Link</a></li><li>Medium. (2024). <em>Beyond the Hype in AI Medical Diagnostics</em>. <a href="https://medium.com/@jessica.moon0102/9cf3089461a3">Link</a></li><li>Click Therapeutics. (n.d.). <em>Revolutionizing care through digital therapeutics</em>. <a href="https://www.clicktherapeutics.com/digital-therapeutics/">Link</a></li><li>Medium. (2024). <em>How about taking a new pill from Digital Therapeutics (DTx)?</em>. <a href="https://medium.com/@hsiehjungting/a42385f60c40">Link</a></li><li>Medium. (2024). <em>Future of the Wearable Landscape in the US: Putting the Power to Take Charge of Your Health in Your Hands</em>. <a href="https://karnika-singh.medium.com/159cf97bae22">Link</a></li><li>Know Labs. (n.d.). <em>BioRFID: A transformative approach to diagnostics</em>. <a href="https://www.knowlabs.co/copy-of-bio-rfid">Link</a></li><li>Lark Health. (n.d.). <em>Artificial Intelligence — AI | Lark Health</em>. <a href="https://www.lark.com/resources/lark-health-ai-artificial-intelligence">Link</a></li><li>BMJ. (2020). <em>Artificial intelligence versus clinicians: systematic review of design, reporting standards, and claims of deep learning studies</em>. <a href="https://www.bmj.com/content/368/bmj.m689">Link</a></li><li>WSJ. (2024). <em>Digital health venture investment rises, but many deals have yet to pay off</em>. <a href="https://www.wsj.com/articles/digital-health-venture-investment-rises-but-many-deals-have-yet-to-pay-off-d51a312f">Link</a></li><li>WeAreCommunity. (n.d.). <em>AI Revolutionizing Healthcare: A Look at IBM Watson’s Impact in the Healthcare Industry</em>. <a href="https://wearecommunity.io/communities/healthcare/articles/5012">Link</a></li><li>DeepMind. (n.d.). <em>Developing reliable AI tools for healthcare</em>. <a href="https://deepmind.google/discover/blog/codoc-developing-reliable-ai-tools-for-healthcare/">Link</a></li><li>Philips. (n.d.). <em>Image-guided therapy powered by AI</em>. <a href="https://www.philips.co.uk/healthcare/e/image-guided-therapy">Link</a></li><li>Health Imaging. (2024). <em>Commercially available AI reduces radiologists’ workload by 34% in certain screening settings</em>. <a href="https://healthimaging.com/topics/artificial-intelligence/commercially-available-ai-reduces-radiologists-workload-34-certain-screening-settings">Link</a></li><li>Fierce Biotech. (2024). <em>Roche introduces AI-powered diabetes tracker to predict blood sugar highs and lows</em>. <a href="https://www.fiercebiotech.com/medtech/roche-introduces-ai-powered-diabetes-tracker-predict-blood-sugar-highs-and-lows">Link</a></li><li>Julien Florkin. (n.d.). <em>AI Fraud Detection: Boost Your Security and Reduce Losses</em>. <a href="https://julienflorkin.com/business/insurance/ai-fraud-detection">Link</a></li><li>AI Multiple. (n.d.). <em>3 Ways AI is Changing the Health Insurance Sector</em>. <a href="https://research.aimultiple.com/ai-health-insurance">Link</a></li><li>InsureTech Connect Asia. (n.d.). <em>Transforming insurance with AI: Sumitomo Life’s journey toward personalized preventive care</em>. <a href="https://asia.insuretechconnect.com/articles/transforming-insurance-ai-sumitomo-lifes-journey-toward-personalized-preventive-care">Link</a></li><li>World Bank. (2023). <em>Ayushman Bharat Digital Mission’s Integrated Digital Health Ecosystem is the Foundation of Universal Citizen-centered Health Care in India</em>. <a href="https://documents1.worldbank.org/curated/en/099081723223028498/pdf/P1750750936a760320bfb9066084a6084bf.pdf">Link</a></li><li>All About AI. (n.d.). <em>19+ Must Know AI Statistics For Healthcare In 2025</em>. <a href="https://www.allaboutai.com/resources/ai-statistics/healthcare/#clinical-and-diagnostic-improvements-resulting-from-ai-in-healthcare">Link</a></li><li>VOA News. (2024). <em>AI enhances maternal healthcare in Kenya</em>. <a href="https://www.voanews.com/a/ai-enhances-maternal-health-care-in-kenya-experts-say/7789467.html">Link</a></li><li>MIT News. (2022). <em>Protecting maternal health with AI in Rwanda</em>. <a href="https://news.mit.edu/2022/protecting-maternal-health-rwanda-0918">Link</a></li><li>Africanews. (2024). <em>Zipline partners with Nigeria for improved healthcare delivery</em>. <a href="https://www.africanews.com/2024/10/25/zipline-partners-nigeria-for-improved-healthcare-delivery">Link</a></li><li>Statista. (n.d.). <em>Artificial intelligence adoption trends in Africa</em>. <a href="https://www.statista.com/outlook/tmo/artificial-intelligence/africa">Link</a></li><li>JAMA Network. (2021). <em>Administrative Expenses in the US Health Care SystemWhy So High?</em>. <a href="https://jamanetwork.com/journals/jama/fullarticle/2785479">Link</a></li><li>HCUP. (2017). <em>Characteristics and Costs of Potentially Preventable Inpatient Stays, 2017.</em> <a href="https://hcup-us.ahrq.gov/reports/statbriefs/sb259-Potentially-Preventable-Hospitalizations-2017.jsp">Link</a></li><li>PubMed. (2024). <em>Medical Error Reduction and Prevention</em>. <a href="https://pubmed.ncbi.nlm.nih.gov/29763131/">Link</a></li><li>CDC. (n.d.). <em>Fast Facts: Health and Economic Costs of Chronic Conditions</em>. <a href="https://www.cdc.gov/chronic-disease/data-research/facts-stats/index.html">Link</a></li><li>Supply Chain Brain. (2024). <em>How to Cut Spending Waste From Healthcare Supply Chains</em>. <a href="https://www.supplychainbrain.com/articles/39143-how-to-cut-spending-waste-from-healthcare-supply-chains">Link</a></li><li>Caresyntax. (n.d.). <em>Making surgery smarter using surgical technology</em>. <a href="https://caresyntax.com/">Link</a></li><li>Smart Reporting. (n.d.). <em>Close your diagnostic reporting gap and alleviate burnout</em>. <a href="https://www.smart-reporting.com/">Link</a></li><li>Resilience. (n.d.). <em>A digital solution dedicated to remote monitoring and support for oncology patients</em>. <a href="https://www.resilience.care/">Link</a></li><li>Tech Funding News. (2024). <em>AI for healthcare: Top funding rounds of Europe in 2024</em>. <a href="https://techfundingnews.com/ai-for-healthcare-top-funding-rounds-of-europe-in-2024/">Link</a></li><li>Mordor Intelligence. (n.d.). <em>Artificial intelligence in healthcare market analysis</em>. <a href="https://www.mordorintelligence.com/industry-reports/artificial-intelligence-in-healthcare-market">Link</a></li><li>Grand View Research. (n.d.). <em>Global outlook on AI in healthcare market size</em>. <a href="https://www.grandviewresearch.com/horizon/outlook/ai-in-healthcare-market-size/global">Link</a></li><li>Deloitte. (2023). <em>From fax machines to gen AI: Are hospitals/health systems ready?</em>. <a href="https://www2.deloitte.com/us/en/blog/health-care-blog/2023/from-fax-machines-to-gen-ai-are-hospitals-health-systems-ready.html?icid=learn-more_click">Link</a></li><li>Deloitte. (2024). <em>Forces of change: the future of healthcare</em>. <a href="https://www2.deloitte.com/us/en/insights/industry/health-care/forces-of-change-health-care.html">Link</a></li><li>Deloitte. (2024). <em>Integrating digital health tools to help improve the whole consumer experience</em>. <a href="https://www2.deloitte.com/us/en/insights/industry/health-care/digital-health-always-on-care.html">Link</a></li><li>Deloitte. (2023). <em>2024 Outlook for Health Care Planning for the Future of Health: Top trends for 2024</em>. <a href="https://www2.deloitte.com/us/en/blog/health-care-blog/2023/outlook-for-health-care.html?id=us:2di:3dr:4outlooks24:eng:chs:120623:cre&amp;icid=learn-more_click">Link</a></li><li>Deloitte. (2023). <em>Overcoming generative AI implementation blind spots in health care</em>. <a href="https://www2.deloitte.com/us/en/insights/industry/health-care/how-to-prepare-for-generative-ai-in-health-care.html">Link</a></li><li>SVB. (2024). <em>Healthcare Industry Trends 2024 Mid-Year Report</em>. <a href="https://www.svb.com/trends-insights/reports/healthcare-investments-and-exits/">Link</a></li><li>BVP. (2024). <em>State of health tech 2024</em>. <a href="https://www.bvp.com/atlas/state-of-health-tech-2024">Link</a></li><li>Sequoia Capital. (n.d.). <em>Generative AI’s Act o1</em>. <a href="https://www.sequoiacap.com/article/generative-ais-act-o1/">Link</a></li><li>Medical Futurist. (2024). <em>The Top 20 Digital Health Trends For The Near Future: Here Is The 2024 Ebook</em>. <a href="https://medicalfuturist.com/top-digital-health-trends/">Link</a></li><li>IDTechEx. (2024). <em>Digital Health and Artificial Intelligence 2024–2034: Trends, Opportunities, and Outlook</em>. <a href="https://www.idtechex.com/en/research-report/digital-health-and-artificial-intelligence-2024-2034-trends-opportunities-and-outlook/985">Link</a></li><li>Medium. (2024). <em>Unlocking Health: Why Patient-Centred Tools are the Next Investment Frontier</em>. <a href="https://medium.com/@drjennywilliams/dce00fc2133a">Link</a></li><li>WHO. (2024). <em>Health workforce: Policies and priorities</em>. <a href="https://www.who.int/health-topics/health-workforce#tab=tab_1">Link</a></li><li>World Economic Forum. (2024). <em>Transforming Healthcare: Navigating Digital Health with a Value-Driven Approach.</em> <a href="https://www3.weforum.org/docs/WEF_Transforming_Healthcare_2024.pdf">Link</a></li><li>McKinsey. (2024). <em>Digital Transformation in Health Systems: Investment Priorities.</em> <a href="https://www.mckinsey.com/industries/healthcare/our-insights/digital-transformation-health-systems-investment-priorities">Link</a></li><li>Binariks. (2024). <em>Artificial Intelligence in Healthcare Market Insights.</em> <a href="https://binariks.com/blog/artificial-intelligence-ai-healthcare-market/">Link</a></li><li>byFounders (2024). <em>The Future of Healthcare</em>. <a href="https://www.byfounders.vc/insights/future-of-healthcare">Link</a></li></ol><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=649abd6a0a96" width="1" height="1" alt=""><hr><p><a href="https://medium.com/included-vc/ai-is-rewriting-the-rules-of-healthcare-a-2024-investors-playbook-649abd6a0a96">AI is Rewriting the Rules of Healthcare: A 2025 Investor’s Playbook</a> was originally published in <a href="https://medium.com/included-vc">Included VC</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Investing in Biotech’s AI Gold Rush: What Smart Money Needs to Know About the Future of Medicine]]></title>
            <link>https://medium.com/included-vc/investing-in-biotechs-ai-gold-rush-what-smart-money-needs-to-know-about-the-future-of-medicine-05d2d9841ee1?source=rss-6bd5914fe82e------2</link>
            <guid isPermaLink="false">https://medium.com/p/05d2d9841ee1</guid>
            <category><![CDATA[biotech]]></category>
            <category><![CDATA[tech]]></category>
            <category><![CDATA[venture-capital]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[pharma]]></category>
            <dc:creator><![CDATA[Ugo Tomasello]]></dc:creator>
            <pubDate>Sat, 14 Dec 2024 21:12:14 GMT</pubDate>
            <atom:updated>2025-01-14T15:46:17.350Z</atom:updated>
            <content:encoded><![CDATA[<h4><strong>Smart Investments in AI-Driven Biotech</strong></h4><h3><strong>Biotech’s AI Gold Rush: What Smart Money needs to learn to invest on the Future of Medicine</strong></h3><p><em>The fusion of AI and biotechnology is transforming drug discovery and reshaping venture capital strategies. Here’s how VCs can capitalize on this high-tech shift in 2024.</em></p><figure><img alt="Illustration of a human head profile composed of geometric, digital patterns, symbolizing AI and biotechnology integration. Two digital, pill-like shapes on the left emit colorful triangular particles, suggesting the concept of drug discovery through AI. The image is surrounded by logos of notable biotech and venture capital companies, including a16z, Kurma Partners, V-Bio Ventures, Forbion, VIB, Insilico Medicine, and insitro, representing key players in the biotech and AI investment landscape." src="https://cdn-images-1.medium.com/max/1024/1*VDXDkeLlJbTXhK2zx1VOCw.png" /><figcaption>The Convergence of AI and Life Sciences: Top Investors and Innovators Shaping the Future of Drug Discovery and Biotechnology</figcaption></figure><p><strong>tl;dr</strong></p><ul><li><strong>AI is reshaping biotech</strong>, speeding up drug discovery and prompting new strategies for venture capital.</li><li><strong>Rising investments reflect AI’s potential</strong> for faster timelines, higher success rates, and scalable models in biotech.</li><li><strong>VCs face key challenges</strong> in regulatory uncertainty, talent gaps, and data quality, all requiring accurate risk management.</li><li><strong>High-impact investment</strong> areas include <strong>more transparent AI models</strong>, scalable tech platforms, and collaborations with pharmaceutical leaders.</li></ul><p><strong>. . .</strong></p><h3>🧭 <strong>Introduction: A new era for Biotech and Venture Capital</strong></h3><p>Artificial Intelligence (AI) is redefining biotech, accelerating drug discovery and introducing fresh approaches for venture capital (VC). As AI’s role expands in 2024, VCs are presented with new avenues for faster, more cost-effective R&amp;D and novel methods for evaluating biotech opportunities. Traditional models with long, high-risk timelines are giving way to AI-driven processes that offer agility and precision. In this article we will explores together the key trends and obstacles in AI-enabled biotech, outlining insights for VCs looking to seize these emerging opportunities.</p><figure><img alt="Treemap showing the amount of new capital flowing into venture funds from 2000 to 2023, measured in billions of dollars. The chart highlights a gradual increase in funding over the years, with significant spikes in 2018 ($14B), 2019 ($29B), 2020 ($31B), and 2021 ($25B), followed by a decrease to $21B in 2023. Early years from 2000 to 2013 show relatively low levels, generally below $10B annually." src="https://cdn-images-1.medium.com/max/1024/1*7uwj8TwHJ1ooDTq7s2ezmA.png" /><figcaption>New Capital Flowing into Venture Funds (2000–2023) investing in biotech industry</figcaption></figure><h4><strong>VC Investments in Biotech (2020–2024): Key Numbers and Trends</strong></h4><p><strong><em>2020–2021: The Pandemic-Fueled Biotech Boom. </em></strong>The COVID-19 crisis fueled unprecedented investments in biotech, reaching $115 billion in 2021. Early-stage companies, especially those focused on cancer and multi-therapy solutions, secured 66% of this capital. AI-focused startups gained VCs’ traction, as the potential to accelerate drug discovery sparked keen interest in predictive models for drug interactions, trial efficiency, and shorter development timelines.</p><p><strong><em>2022: Biotech Funding Adjusts, AI Thrives</em></strong>. After the pandemic, biotech funding adjusted to market shifts, with $16.3 billion raised by Q1. Despite this cooling, AI-centered startups continued to capture attention, securing $500 million in Series A funding, reflecting steady interest in AI applications. Investors began favoring later-stage, lower-risk ventures, a trend extending into 2023.</p><p><strong><em>2023–2024: Stable but Selective Investments. </em></strong>Biotech funding settled, with $21.5 billion raised in 2023, marking a post-pandemic equilibrium. Startups using AI for drug discovery, molecule design, and gene editing maintained strong appeal. Licensing deals for AI-based drug technologies held their value, signaling sustained confidence in AI’s impact on biotech.</p><h3>⏭️ <strong>The Role of AI in Reshaping VC in Biotech</strong></h3><h4><strong><em>Accelerating Drug Discovery: The AI Edge</em></strong></h4><p>Traditional drug discovery requires years of costly screening and refinement, taking 4–6 years and $150–$500 million to find viable candidates. AI is reshaping this approach by analyzing extensive datasets, identifying complex patterns, and simulating biological interactions in-silico, predicting successful drug candidates before lab work begins. This approach promises faster, cheaper drug discovery with fewer failures. Beyond managing large volumes of data, AI is able to recognize subtle patterns of information and manipulate them to approach biological questions under a new perspective. Synthetic biology, in particular, stands to gain significantly from these AI-driven perspectives.</p><h4><strong><em>AI Models Transforming Drug Discovery</em></strong></h4><p>Advanced AI models like Convolutional Neural Networks (CNNs), Generative Adversarial Networks (GANs), and Reinforcement Learning (RL) are transforming drug discovery.</p><p>· <a href="https://www.mdpi.com/1420-3049/29/19/4626"><strong>CNNs</strong></a><strong> </strong>predict protein and RNA structures, improving the understanding of drug-target interactions.</p><p>· <a href="https://www.mdpi.com/1420-3049/29/19/4626"><strong>GANs</strong></a> generate novel protein sequences for synthetic biology, speeding de novo drug design.</p><p>· <a href="https://www.mdpi.com/1420-3049/29/19/4626"><strong>RL</strong></a> optimizes protein stability and activity, helping researchers explore mutation spaces for therapeutic advances.</p><figure><img alt="This image illustrates four aspects of Explainable AI (XAI): Explain to Justify (Red) — Represented by sliders, showing AI’s ability to justify decisions. Explain to Control (Purple) — A hand icon symbolizes enabling human control. Explain to Discover (Yellow) — A lightbulb highlights aiding discovery through insights. Explain to Improve (Gold) — Arrows signify refining AI systems. Vibrant colors and icons emphasize interpretability and transparency in AI." src="https://cdn-images-1.medium.com/max/700/1*zymMdeV1iKeqzNkeX8CZIQ.jpeg" /><figcaption><a href="https://www.birlasoft.com/articles/demystifying-explainable-artificial-intelligence">The Four Pillars of Explainable AI (XAI): Justify decisions, enable control, foster discovery, and drive continuous improvement in AI systems.</a></figcaption></figure><p>To dive deeper on AI, I suggest to read <a href="https://medium.com/included-vc/beyond-the-hype-in-ai-medical-diagnostics-9cf3089461a3"><strong>Beyond the Hype in AI Medical Diagnostics</strong></a> Medium article, written by <a href="https://www.linkedin.com/in/jessmoon-phd/">Jessica Moon</a>.</p><h4>🤖 <strong><em>Building Trust in AI with Explainable AI (XAI)</em></strong></h4><p>To bring some clarity on the real of AI models and their adoption, I interviewed <strong>Prof.</strong> <strong>Pedro Gonçalves</strong>, Group Leader at <a href="https://nerf.be/en#/">Nerf</a> and Associate Professor at <a href="https://wms.cs.kuleuven.be/cs/english">KU Leuven</a> Departments of Computer Science and Electrical Engineering.</p><p><strong><em>Why we need AI?</em></strong></p><p>One of the aspect of the current technological era is the vast amount of data generated and collected. Both at academic research and industrial application, humans struggle to extract meaning from the information acquired. It’s here that AI come in play, to find the relevant information and frame it in a meaningful structured output.</p><p><strong><em>Which features of AI can support Biotech innovation?</em></strong></p><p>In the Biotech field you want to generate outputs that are very accurate and trustable, as they might impact the health and life of human being. The AI model need to be very robust, while dealing with the perturbations embedded in the biological variability. A key feature is the use of <strong>Explainable AI (XAI) vs the “black box” models</strong>. While “black box” models work best when the output can be trusted without further understanding, the use of XAI help to understand the rationale behind AI-generated predictions, improving transparency in drug discovery​. The problem with XAI is that there are not enough models out there yet to outcompete the “black box” ones. A complementary approach to XAI is the Informed AI, which allows to integrate domain-specific human knowledge with AI model, so that the output does not exclusively rely on database. The use of these hybrid model might be very useful to predict key preclinical parameters such as pharmacokinetics and pharmacodynamics, as we already developed a good understanding of them.</p><p><strong><em>So, do you think Biotech will be adopting XAI on a larger scale?</em></strong></p><p>XAI can bring a better understanding of biological processes that we still don’t understand very well. The acquisition of this knowledge may be very valuable and provide a competitive advantage in drug development.</p><h4><strong>📜 <em>Regulatory Backing for XAI</em></strong></h4><p>Regulatory bodies are recognizing AI’s value. In 2023, the FDA began promoting a flexible, risk-based framework for AI, aimed at accelerating drug approvals while ensuring safety. Trustworthy AI, especially XAI, could be a critical enabler for regulatory acceptance of AI-driven drug discoveries.</p><blockquote>The problem with XAI is that there are not enough models out there yet to outcompete the “black box” ones.</blockquote><h3>🧪<strong><em>AI-Driven Startups Capturing VC Interest</em></strong></h3><p>AI advancements in drug discovery have positioned biotech startups as attractive prospects for VCs. Companies like Insilico Medicine and Insitro are leading the way, using AI to cut time and costs in drug development.</p><p>· <a href="https://insilico.com/"><strong>Insilico Medicine</strong></a><strong>:</strong> Known for ISM001–055, the first drug designed entirely by AI, developed in under 18 months — a significant achievement compared to the typical 5–10 years. Their Pharma.AI framework include several outstanding platforms. Insilico’s PandaOmics platform identifies biological targets, while Chemistry42 creates optimized structures. With over $400 million raised and 30 AI-designed drugs in its pipeline, Insilico’s rapid success has set a new standard in Biotech.</p><figure><img alt="This image depicts Pharma.AI’s integrated drug discovery system powered by generative AI and robotics. It features three platforms: Biology42, Chemistry42, and Medicine42, each with dedicated applications like pandaOmics, Generative Chemistry, and inClinico. The visual highlights interconnected systems using circuit-like graphics, symbolizing seamless integration across biology, chemistry, and clinical solutions." src="https://cdn-images-1.medium.com/max/1024/1*kTMEJ-mm89EAshN9BqspLg.jpeg" /><figcaption><a href="https://pharma.ai/">Pharma.AI: Revolutionizing Drug Discovery with Fully-Integrated Generative AI and Robotics</a></figcaption></figure><p>· <a href="https://www.insitro.com/"><strong>Insitro</strong></a><strong>:</strong> Led by Daphne Koller, Insitro’s platform combines AI and robotic automation to study diseases through induced pluripotent stem cells (iPSCs), offering deep insights into disease mechanisms. Their approach has attracted $643 million in funding, advancing the potential for AI to tackle complex diseases.</p><h4><strong>📈 <em>Investing in AI-Driven Biotech: Key Insights and Evaluation Criteria</em></strong></h4><p>To understand how AI is transforming biotech and attracting VC’s interest, I interviewed <strong>Dennis Pedri</strong>, Analyst at <a href="https://v-bio.ventures/">V-Bio Ventures</a>, a VC focusing on early-stage therapeutics across Europe and funding spin-off companies from <a href="https://vib.be/en#/"><strong>VIB</strong></a>, a leading Biotech research institute. I also interviewed <strong>Caroline Goddeeris</strong>, a Partner at a Belgian VC fund specializing in therapeutics. Their first-hand experiences shed light on the criteria and challenges they consider when evaluating AI-driven biotech ventures.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/502/1*lBC3IP4yOrnuJSDxzAyamA.png" /></figure><p><strong><em>How do you approach evaluating cutting-edge biotech companies for investment?</em></strong></p><p><strong>Dennis</strong>: Apart the set of standard criteria that most of investment funds look into (company stage of development, competition, market size, etc.), the main question is whether the company or the technology can fill a gap or if there is an unmet need that it is trying to solve. If the company has a strong societal impact, it will likely be a profitable investment.</p><p><strong>Caroline</strong>: Especially for the early-stage ventures, there is often a higher risk, considering technology validation, translation questions, manufacturing and maturity of the field, etc. You don’t want to add risk on risk. For example, some years ago we have been evaluating a company developing Antibody-Drug-Conjugates (ADCs), and the risks we identified were associated with the scalability of the manufacturing process, as it wasn’t a Plug &amp; Play process into any Contract Development and Manufacturing Organization (CDMO), and the ADC risk itself, as this class of molecules were not validated at that time.</p><p><strong><em>How well you need to understand the technology that is behind the venture, to decide to invest?</em></strong></p><p><strong>Dennis</strong>: We aim for a deep understanding, analyzing the biology, technology, and application field. Expertise, Key Opinion Leaders (KOLs), and networking with other funds inform our due diligence.</p><p><strong><em>How do you get familiar with the use of AI in Biotech? Do you already have some company in your portfolio that use AI as core of their technology?</em></strong></p><p><strong>Dennis</strong>: I’m seeing in the Biotech market two big trends regarding the usage of AI. One is diagnostics and the other one is drug discovery and development. For the latter, we have, for example, a strong positive example with one of our portfolio companies, called Sibylla Biotech. It is an Italian company, exploiting AI computer simulations to develop small molecules to induce the degradation of a target by interfering with its folding pathway. To do that, the model simulates numerous intermediate folding states and how they can be exposed to ligand compounds. When we evaluated Sibylla, we certainly took into consideration the potential of their AI platform, as it is instrumental for the company itself.</p><p><strong>Caroline</strong>: So far, I’ve observed AI being used in a somewhat scattered way throughout the drug development process. As for drug discovery, it is often an extension of existing programs already common in labs, now enhanced with an additional dataset layer that can be applied to drug discovery with a specific focus. The main question, however, remains: how insightful, robust and complete is your dataset, and how unique is it to your company? We’re seeing more drug development companies saying, “Okay, it’s AI-supported,” and presenting their datasets, but the question remains: how should we evaluate that? I’m still figuring that out. So yes, it often comes back to the datasets, which is now our focus, making sure the data is valuable, differentiated, and ideally unique in some way. As a Biotech-VC, we primarily evaluate ventures based on the dataset, especially the preclinical data. It’s a different kind of input, but it still comes down to the quality of the data they bring, regardless of how they arrived at it.</p><p><strong><em>Do you think that the use of AI, to speed up the Drug Development cycle or to be the core technology of a Biotech platform, will bring to smaller and flexible investments?</em></strong></p><p><strong>Dennis</strong>: It’s a possibility. It’s still something that needs to be proven during the process of a specific deal. When we evaluate a company, at the later stages of the due diligence, we also look at how the budget of the investment is going to be spent. If the company shows convincingly that they can reach their objectives with a smaller investment thanks to their AI based methods, it is certainly an added value that increase the chances of the deal.</p><blockquote>“It’s a different kind of input, but it still comes down to the quality of the data they bring, regardless of how they arrived at it.”</blockquote><h4><strong><em>AI’s Appeal to Venture Capitalists</em></strong></h4><p>The escalating costs and complexities of drug development have compelled the biotech industry to seek innovative solutions. Artificial intelligence (AI) has emerged as a transformative force, attracting significant venture capital (VC) investment due to its potential to enhance efficiency and success rates in drug discovery and development.</p><p><strong>Resource Optimization</strong></p><ul><li><a href="https://hitconsultant.net/2024/11/13/nventures-invests-in-genesis-therapeutics-ai-platform-for-drug-development/?utm"><strong>Prioritizing Promising Candidates</strong></a>: AI algorithms analyze vast datasets to identify and prioritize drug candidates with the highest potential, reducing the time and resources spent on less viable options. For instance, <a href="https://genesistherapeutics.ai/"><strong>Genesis Therapeutics</strong></a> utilizes its proprietary AI platform, GEMS, to revolutionize small molecule drug discovery, attracting substantial VC funding.</li><li><a href="https://techcrunch.com/2021/06/22/a-i-drug-discovery-platform-insilico-medicine-announces-255-million-in-series-c-funding/?utm"><strong>Reducing Laboratory Expenses</strong></a>: By simulating biological processes and predicting molecular interactions, AI minimizes the need for extensive laboratory experiments, thereby cutting costs. Companies like <a href="https://insilico.com/"><strong>Insilico Medicine</strong></a> have raised significant capital to develop AI-driven platforms that streamline the drug discovery process.</li></ul><p><strong>Enhancing Clinical Success</strong></p><ul><li><a href="https://www.mckinsey.com/industries/life-sciences/our-insights/generative-ai-in-the-pharmaceutical-industry-moving-from-hype-to-reality?utm"><strong>Improved Candidate Selection</strong></a>: AI-driven predictions assist in selecting drug candidates with a higher likelihood of success in clinical trials, thereby improving overall success rates. McKinsey reports that generative AI can lead to up to 50% cost reductions by streamlining clinical trial processes and auto-drafting trial documents.</li><li><a href="https://www.reuters.com/technology/artificial-intelligence/nvidia-backed-ai-firm-iambic-unveils-drug-discovery-breakthrough-2024-10-29/?utm"><strong>Accelerating Development Timelines</strong></a>: AI models can predict drug impacts on the human system earlier, reducing late-stage failures and accelerating the development process. For example,<strong> </strong><a href="https://www.iambic.ai/"><strong>Iambic Therapeutics</strong></a>’ AI model, Enchant, has demonstrated the potential to significantly cut down the time and cost required for drug development.</li></ul><p><strong>Addressing Unmet Needs</strong></p><ul><li><a href="https://techcrunch.com/2024/04/24/xaira-an-ai-drug-discovery-startup-launches-with-a-massive-1b-says-its-ready-to-start-developing-drugs/?utm"><strong>Exploring New Biological Targets</strong></a>: AI enables the exploration of complex biological targets and pathways, facilitating the development of treatments for diseases that were previously difficult to address. Companies like <a href="https://www.linkedin.com/company/xaira-therapeutics/"><strong>Xaira Therapeutics</strong></a> have been funded to leverage AI in developing novel therapies for such conditions.</li><li><a href="https://www.thetimes.com/business-money/technology/article/ai-new-drugs-and-the-appliance-of-life-science-p8q0sw8j0?utm"><strong>Opportunities in Rare Diseases</strong></a>: AI’s ability to analyze complex datasets allows for the identification of potential treatments for rare diseases, a field often overlooked due to limited financial incentives. Startups like <a href="https://healx.ai/"><strong>Healx</strong></a> are utilizing AI to find treatments for rare diseases, attracting VC interest.</li></ul><h4>European VC’s Role in the AI-Driven Biotech Boom</h4><p>European VC funds are increasingly recognizing the potential of AI in biotech, particularly in accelerating drug discovery and therapeutic development. Leading firms such as Sofinnova Partners and Kurma Partners are driving investments in this area. <a href="https://sofinnovapartners.com/"><strong>Sofinnova</strong></a> has invested in <a href="https://www.kiro.bio/solution?lang=en">Kiro</a>, a company using AI to streamline lab data interpretation, advancing precision medicine within clinical settings. <a href="https://www.kurmapartners.com/en"><strong>Kurma Partners</strong></a> backs <a href="https://memo-therapeutics.com/dropzylla/#approach">Memo Therapeutics</a>, which applies computational methods to streamline antibody discovery, aiming to reduce timelines for biologics development.</p><p><a href="https://www.bpifrance.fr/"><strong>Bpifrance</strong></a><strong> </strong>has also shown its commitment by funding <a href="https://www.bioptimus.com/about">Bioptimus</a>, whose AI model aids cancer detection through advanced data analytics. <a href="https://earlybird.com/"><strong>Earlybird Venture Capital</strong></a> and <a href="https://forbion.com/en"><strong>Forbion</strong></a> are similarly active, investing in AI-enabled biotechs with tools like Earlybird’s Eagle Eye platform for pipeline optimization and Forbion’s <a href="https://airna.com/our-science/">AIRNA</a>, which developed a platform to modify protein functionality towards RNA machinery . This landscape underscores Europe’s growing commitment to fostering cutting-edge biotech progress at the intersection of AI and therapeutics.</p><figure><img alt="Collection of logos from various venture capital firms focused on biotech and healthcare. Includes logos of Mercia Fund Managers, Earlybird, AlbionVC, Sofinnova Partners, NVF (Novartis Venture Fund), Kurma Partners, O2H Ventures, Forbion Ventures, Novo Holdings, Apollo Health Ventures, Bpifrance, and V-Bio Ventures. Created by Ugo Tomasello and associated with Included VC" src="https://cdn-images-1.medium.com/max/1024/1*FmhYoAebWEMN8ucv8n8I2w.png" /><figcaption><strong>Prominent Life Sciences and Healthcare Venture Capital Funds: Driving innovation in biotech, medtech, and health solutions globally</strong></figcaption></figure><h4>🍊 <strong><em>Investment Perspective from a16z </em></strong>🍊</h4><p>Vineeta Agarwala of <a href="https://a16z.com/"><strong>Andreessen Horowitz</strong></a> (a16z) highlights two main VC value drivers for AI: <strong>Efficiency</strong> gains and enhanced Efficacy/Safety. AI shortens the time-to-IND (Investigational New Drug), allowing companies to target more projects with the same capital, reducing risk. Key metrics time to IND, probability of success (PTS), and Capex per IND are crucial to improve efficiency.</p><p>By integrating <strong>Efficacy and Safety</strong> standards in AI design, companies can de-risk earlier, increasing clinical trial success and minimizing capital needs. AI is also unlocking new strategies for identifying <strong>hard-to-detect targets</strong> and binding pockets, leading to the development of novel molecules or innovative methods of delivery to the target.</p><h3>💸 <strong>AI in Biotech: Investment Opportunities and Challenges</strong></h3><h4>🔑 <strong><em>Key Opportunities for VCs</em></strong></h4><p>The integration of AI into Biotech brings new opportunities for VC.</p><ul><li>By <strong>accelerating drug discovery timelines</strong> and reducing associated costs, AI enables startups to reach milestones with less capital. This increased pace and resource efficiency make AI-driven biotech companies attractive to investors, as they can advance therapies from early development to clinical trials with notable agility. Additionally, the precision that AI offers has the potential to improve clinical trial outcomes, reducing some of the risks associated with biotech investments.</li><li>For VCs, the <strong>scalability of AI platforms</strong> adds further value, as many platforms can be applied across various therapeutic areas, broadening the reach of a single investment. The therapeutic widening application allows AI platforms for drug discovery to strongly leverage on a Software-as-a-Service (SaaS) or licensing model, creating recurring revenue streams.</li><li><strong>Strategic alliances with pharmaceutical giants</strong> also strengthen these opportunities. Large Pharma companies, eager for assets that offer a competitive edge, are increasingly partnering with AI-driven biotech startups. These partnerships grant AI-focused companies access to extensive clinical data, accelerating their development cycles. In the past five years, the top 20 pharmaceutical companies have formed over 55 partnerships in AI, with initial payments exceeding $770 million and potential future value over $38 billion.</li><li>The <strong>expansion of personalized medicine</strong> offers additional paths for AI-driven biotech to address unmet needs in specific patient populations, potentially lowering healthcare costs. Personalized drug design allows for treatments with improved efficacy and fewer side effects, which can increase the chances of regulatory approval and market access, paving the way for revenue generation.</li><li><strong>Drug repurposing</strong>. Investing in startups building AI platforms capable of screening thousands of compounds simultaneously for repurposing potential represent a very interesting opportunity. This can be build on <a href="https://arxiv.org/abs/2303.00240?utm_source"><strong>ML</strong></a>, where the algorithm identifies off-target effects assessing drug-target interactions, <a href="https://bmcbiol.biomedcentral.com/articles/10.1186/s12915-024-02028-3?utm_source"><strong>Natural Language Processing (NLP)</strong></a>, where AI scan scientific literature clinical trials, and patient data to uncover hidden links between drugs and diseases, and <a href="https://link.springer.com/article/10.1007/s00044-024-03253-9?utm_source"><strong>Molecular Docking Simulations</strong></a>, where AI models simulate how existing drugs bind to new targets, accelerating hypothesis testing. The drug repurposing market is projected to grow at a CAGR of ~8%, expected to exceed $35 billion by 2028.</li></ul><figure><img alt="A circular diagram highlights five aspects of life sciences innovation: Personalized Medicine (DNA icon), Drug Repurposing (recycling icon), Accelerate Discovery (rocket icon), Platform Scalability (bar chart), and Strategic Alliances (puzzle pieces). A star at the center symbolizes their synergy, with colorful brushstroke backgrounds. Included VC and creator Ugo Tomasello are credited at the bottom left." src="https://cdn-images-1.medium.com/max/1024/1*FwGxVwwvC95E4SU--UQpDA.png" /><figcaption>Investment Opportunities for VCs in AI-driven biotechs</figcaption></figure><blockquote>In the past five years, the top 20 pharmaceutical companies have formed over 55 partnerships in AI, with initial payments exceeding $770 million and potential future value over $38 billion.</blockquote><h4>🚧 <strong><em>Key Challenges for VCs</em></strong></h4><p>While AI-driven biotech offers substantial potential, several challenges remain. <strong>AI adoption across the drug development process remains fragmented</strong>, creating a heterogeneous landscape that complicates the evaluation of competitive advantages. Furthermore, the effectiveness of AI models depends on high-quality data. In fields where data is scarce or inconsistent, like rare diseases, there’s a higher risk of inaccurate predictions and missed insights. The “<strong>black box</strong>” nature of some AI models also brings regulatory challenges, as regulatory agencies increasingly seek transparency in how AI generates its predictions. Another significant obstacle is the shortage of talent with expertise in both AI and biotech, which raises recruitment costs and can lead to turnover, impacting project stability.</p><p>For VCs, investing in AI-driven Biotech holds great promise, but it demands a rigorous approach to evaluate the quality of data, model transparency, and team expertise. Success in this space will require <strong>building strong</strong> <strong>partnerships, prioritizing explainable AI</strong>, and establishing robust talent retention strategies, ensuring that these investments can deliver sustainable value in the rapidly evolving biotech landscape.</p><p>Next, we’ll take a closer look at the heavy reliance on data, the shortage of skilled talent, and the critical role of strategic partnerships before summarizing our key takeaways.</p><figure><img alt="This image contrasts two types of AI: “Black Box AI,” represented by a dark box, signifies opaque methods like Convolutional Neural Networks, Generative Adversarial Networks, and Reinforcement Learning. A figure attempts to open the box, symbolizing efforts to understand its inner workings. Opposite, “Explainable AI” is depicted with a simple icon of a human head and neural connections, highlighting transparency and interpretability. The design emphasizes the need for clarity in AI systems." src="https://cdn-images-1.medium.com/max/1024/1*wICP5DD-zMErnMCquenX4A.png" /><figcaption><a href="https://umdearborn.edu/news/ais-mysterious-black-box-problem-explained">Contrasting AI Approaches: Black Box vs Explainable AI</a></figcaption></figure><h4>🤝 <strong><em>The High-Stakes Equation: Data, Talent, and Partnerships in AI-Driven Biotech</em></strong></h4><p>AI-driven Biotech companies rely heavily on large, high-quality datasets to train models and generate accurate predictions. In drug discovery, a model trained on biased or incomplete data might incorrectly predict a drug’s efficacy or overlook potential side effects, leading to costly setbacks in later stages. This challenge is particularly relevant in fields like rare diseases, where data scarcity is common. For VCs investing in biotech companies targeting rare diseases, data limitations pose a real risk to development timelines. To mitigate the risk, the use of <strong>Transfer Learning Models</strong> allows gain of knowledge from large datasets to enhance the performance on limited data.</p><p>However, accessing these data sources can be expensive, especially when licensing agreements are needed with hospitals or research institutions. This piles up with significant costs involved in standardizing, cleaning, and preparing data for ML models. Understanding data source and database integrity is crucial for VCs to estimate the impact on the company operation expenses and potential. <strong>In-house data generation</strong> provides an alternative that can provide a self-sustaining data pipeline and creates valuable assets. Proprietary datasets allow companies to avoid external data costs and they can also open up new revenue streams through data licensing or form the basis for strategic partnerships. This approach can strengthen a company’s technological and financial position over time.</p><p>The <strong>shortage of skilled talent </strong>in AI-driven biotech brings additional operational hurdles. Difficulty in recruiting and retaining qualified experts leads to project delays, extending timelines for investment returns and impacting VC fund cycles. Competitive salaries and recruitment cost strain early-stage budgets, often requiring additional capital from investors. Heavy reliance on a small number of specialized individuals also increases the risk, as it creates a <strong>single point of failure</strong>. Scaling without sufficient skilled personnel stalls development and delays time-to-market, directly affecting returns. To address these challenges, investors might consider setting aside funds for workforce development to support talent building within companies. Strategic partnerships with universities and tech firms can help bridge skill gaps, while remote work options expand access to global AI expertise.</p><figure><img alt="This image outlines challenges in AI-driven drug discovery across three areas: Data, Tools, and Capabilities. Key issues include limited data maturity and access, lack of advanced tools, and insufficient expertise and infrastructure, all impacting AI adoption in drug development." src="https://cdn-images-1.medium.com/max/1024/1*ULDOnSkr_979KrC7XGcoDQ.png" /><figcaption>Key Challenges in AI-Driven Drug Discovery: Addressing gaps in data maturity, tool accessibility, and standardization to unlock the full potential of artificial intelligence in healthcare innovation</figcaption></figure><p><strong>Pharmaceutical partnerships with AI-driven biotech firms</strong> require close alignment between AI models, data, and the therapeutic priorities of the Pharma companies. This alignment is essential, as large pharmaceutical companies often face obstacles in adapting AI within established R&amp;D structures. Many are testing AI’s potential through collaborations with ‘AI-first’ biotech firms, while others are making substantial investments to build in-house AI capabilities.</p><h4>🔥 The $500 billion Pharma Firepower 🔥</h4><p>A recent CB Insights study ranked major Pharma companies on their AI readiness across three metrics: <strong>Talent, Execution and Innovation.</strong> hese rankings were based on factors like patent activity, partnership and licensing deals, acquisitions, key hires, product launches, and earnings reports. Roche, Bayer, and Novartis emerged as leaders, actively using AI to improve drug discovery and advance drugs from lab research to clinical trials. <strong>Notably, Roche and Novartis lead among the 16 top Pharma companies with significant liquidity (often referred to as “firepower”), collectively totaling around $500 billion.</strong> AstraZeneca’s strategic partnership with BenevolentAI exemplifies the industry’s expanding AI collaborations, with five new drug targets added to AstraZeneca’s pipeline as a result.</p><p>In 2023, however, there was a notable shift in Pharma M&amp;A towards clinical-stage biotechs, indicating some risk-aversion toward earlier, preclinical stages (<strong>$24 billion in clinical-stage investments in 2023 compared to $14 billion in 2022</strong>). Despite this trend, AI-first biotechs and their investors are largely focused on certain therapeutic areas and molecule types. <strong>Small molecules account for about 90%</strong> of the more than 400 therapeutic pipelines in AI-focused biotech, with a strong emphasis on data-rich fields like <strong>oncology</strong> and <strong>neurology</strong>. As AI-driven drug candidates progress through clinical trials and regulatory approvals, they are likely to gain further trust from pharmaceutical partners, even in the preclinical phase.</p><p>For investors, collaborating with partners connected to the pharmaceutical sector can accelerate strategic alliances and increase the likelihood of success. These connections support smoother negotiations, clearer alignment on objectives, and ultimately increase the potential for translating AI-driven therapies into market-ready treatments.</p><figure><img alt="Chart evaluating pharmaceutical companies across three categories: Innovation, Execution, and Talent. Companies like Roche, Bayer, and Merck score in Innovation; AstraZeneca, Sanofi, and Amgen in Execution; and Moderna, Novartis, and Johnson &amp; Johnson in Talent. Star ratings indicate performance in each category. Created by Ugo Tomasello, associated with Included VC" src="https://cdn-images-1.medium.com/max/1024/1*bu9yPzayT1NGFMsTEXK_zw.png" /><figcaption><a href="https://www.cbinsights.com/research/ai-readiness-index-pharma/">Key Players in Biopharma: Evaluating Leading Companies Across Innovation, Execution, and Talent Capabilities</a></figcaption></figure><h3>⚖️ <strong>Conclusions</strong></h3><p>Reflecting on the role of AI in biotech, I am struck by both its vast potential and the careful stewardship it demands from investors. The intersection of these fields offers VCs a unique chance to accelerate drug discovery, reduce development timelines, and improve patient outcomes. However, to realize these benefits fully, investments must strategically address three essential areas: Data, Models, and Talent.</p><p><strong>Data remains the fuel</strong> of successful AI-driven Biotechbut often fragmented, especially in niche fields like rare disease research. VCs can make a significant impact by supporting data-sharing efforts, strategic partnerships, and <strong>standardization frameworks</strong> like FAIR (Findable, Accessible, Interoperable, and Reusable) and PRIDE (PRoteomics IDEntification) to ensure data quality. Evaluating a company’s ability to generate proprietary data is also essential.</p><p><strong>AI Tools</strong> must progress in accessibility and transparency. Investments in XAI platforms are essential to meet regulatory requirements and foster trust, while open-source initiatives can broaden access, allowing innovation to thrive across biotech.</p><p><strong>Talent</strong> remains the cornerstone of growth. To bridge the skills gap, VCs can foster workforce development through training programs and partnerships with academic institutions.</p><p>In essence, the road to realizing AI’s full potential in biotech requires strategic, multidimensional investment. VCs who prioritize these areas — data quality, tool accessibility, and talent development — are poised to lead the AI-driven biotech revolution, advancing therapies that deliver real-world impact while achieving sustainable returns.</p><p>I hope this article has shed some light on the practical steps and strategic considerations needed to drive meaningful progress, showing how, with the right approach, AI can foster both impactful and profitable advancements in drug discovery and therapeutic investments.</p><blockquote>Let me know in the comments if I miss some key aspect, ventures or investments, I might include them in my next piece!</blockquote><p>#VentureCapital, #Startups, #Investment #AI, #Biotech</p><p><strong>. . .</strong></p><h3>🧩 About me</h3><p>With a background spanning <strong>R&amp;D in Molecular Neurobiology</strong>, Clinical Development Management, and <strong>Venture Capital</strong>, I love to explore new perspectives to evaluate Tech innovation and <strong>investments in Life Sciences</strong>. My commitment to transforming scientific insights into tangible healthcare solutions fuels my passion for this evolving AI-Biotech intersection.</p><p>If you want to discuss these exciting trends further — please feel free to reach out on LinkedIn or via <a href="http://ugo.tomasello@gmail.com/">email</a>.</p><p>👉 <strong>You can find me on </strong><a href="https://www.linkedin.com/in/ugo-tomasello/"><strong>LinkedIn</strong></a><strong> and keep up to date with my latest insights on </strong><a href="https://medium.com/@ugotomasello"><strong>Medium</strong></a></p><p>👉🏼 <strong><em>Follow </em></strong><a href="https://included.vc/"><strong><em>Included VC</em></strong></a><strong><em> on </em></strong><a href="https://twitter.com/included_vc"><strong><em>Twitter</em></strong></a><strong><em> &amp; </em></strong><a href="https://www.linkedin.com/company/includedvc/"><strong><em>LinkedIn</em></strong></a></p><p><strong>#includedvcmafia</strong></p><p><strong>. . .</strong></p><p>🙏 <strong>Special thanks</strong></p><p>A special thank you note to everyone who contributed to the development of this article (the Included VC team, Included VC Delegates, and all the VCs whose work and testimonies allowed me to develop extensive research on the topic). Namely, I would like to thank <a href="https://www.linkedin.com/in/pedro-goncalves-61335b61/">Pedro Gonçalves</a>, <a href="https://www.linkedin.com/in/dennis-pedri-2959ab78/">Dennis Pedri</a>, <a href="https://www.linkedin.com/in/caroline-goddeeris-b83a6b3/">Caroline Goddeeris</a> and <a href="https://www.linkedin.com/in/nikitathakrar/">Nikita Thakrar</a>.</p><p><strong>References</strong></p><p>1. <a href="https://www.mckinsey.com/industries/life-sciences/our-insights/it-will-be-a-paradigm-shift-daphne-koller-on-machine-learning-in-drug-discovery">Koller, D. It will be a paradigm shift: Daphne Koller on machine learning in drug discovery. McKinsey &amp; Company.</a></p><p>2. <a href="https://www.mckinsey.com/industries/life-sciences/our-insights/vineeta-agarwala-on-the-promise-and-limits-of-ai-in-drug-discovery">Agarwala, V. Vineeta Agarwala on the promise and limits of AI in drug discovery. McKinsey &amp; Company.</a></p><p>3. <a href="https://www.mckinsey.com/industries/life-sciences/our-insights/what-early-stage-investing-reveals-about-biotech-innovation">McKinsey &amp; Company. What early-stage investing reveals about biotech innovation.</a></p><p>4. <a href="https://www.marketsandmarkets.com/Market-Reports/ai-in-drug-discovery-market-151193446.html">MarketsandMarkets. AI in drug discovery market.</a></p><p>5. <a href="https://www.ropesgray.com/en/insights/alerts/2024/03/life-sciences-2024-dealmaking-trends-and-outlook">Ropes &amp; Gray LLP. (2024, March). Life sciences 2024 dealmaking trends and outlook.</a></p><p>6. <a href="https://web-assets.bcg.com/86/e5/19d29e2246c7935e179db8257dd5/unlocking-the-potential-of-ai-in-drug-discovery-vf.pdf">Boston Consulting Group. Unlocking the potential of AI in drug discovery.</a></p><p>7. <a href="https://www.jpmorgan.com/content/dam/jpmorgan/documents/cb/insights/outlook/jpmorgan-dec-2023-biopharma-licensing-and-venture-report.pdf">J.P. Morgan. (2023, December). Biopharma licensing and venture report.</a></p><p>8. <a href="https://www.jpmorgan.com/content/dam/jpmorgan/documents/cb/insights/outlook/jpm-biopharma-deck-q2-2024-final-ada.pdf">J.P. Morgan. (2024, Q2). Biopharma deck Q2 2024.</a></p><p>9. <a href="https://www.stifel.com/newsletters/investmentbanking/bal/marketing/healthcare/biopharma_timopler/StifelBiopharmaOutlook2024_01.05.2024.pdf">Stifel Financial Corp. (2024, January 5). Biopharma outlook 2024.</a></p><p>10. <a href="https://www.globenewswire.com/news-release/2022/11/08/2551039/31533/en/Insilico-Medicine-Signs-Strategic-Research-Collaboration-with-Sanofi-worth-up-to-1-2-Billion.html">GlobeNewswire. (2022, November 8). Insilico Medicine signs strategic research collaboration with Sanofi worth up to $1.2 billion.</a></p><p>11. <a href="https://www.ey.com/en_us/life-sciences/biotech-outlook#form">EY. Biotech outlook 2024.</a></p><p>12. <a href="https://wired.me/science/insilico-ai-drug/">O’Brien, M. The company using AI to design drugs faster and cheaper. Wired Middle East.</a></p><p>13. <a href="https://www.labiotech.eu/best-biotech/european-biotech-companies/">LaBiotech. (n.d.). Top European biotech companies.</a></p><p>14. <a href="https://biopharmaapac.com/news/16/4135/2024-biopharma-outlook-24-trends-innovations-and-fdas-critical-role.html">Biopharma APAC. (2024). 2024 Biopharma outlook: 24 trends, innovations, and FDA’s critical role.</a></p><p>15. <a href="https://www.technologynetworks.com/drug-discovery/articles/top-trends-shaping-pharma-in-2024-ai-gene-editing-biosimilars-and-real-world-data-386833">Technology Networks. (2024). Top trends shaping pharma in 2024: AI, gene editing, biosimilars, and real-world data.</a></p><p>16. <a href="https://sigma.software/about/media/top-7-biotechnology-trends-to-watch-in-2024">Sigma Software. (2024). Top 7 biotechnology trends to watch in 2024.</a></p><p>17. <a href="https://insilico.com/blog/first_phase2">Insilico. First phase 2 results for AI drug development.</a></p><p>18. <a href="https://www.casgevy.com/">CASG. CASG solutions.</a></p><p>19. <a href="https://www.mdpi.com/1420-3049/29/19/4626">MDPI. (2024). AI-driven drug discovery: A review. Molecules, 29(19), 4626.</a></p><p>20. <a href="https://www.linkedin.com/pulse/designer-drugs-20-ai-ushers-new-era-drug-discovery-dr-jonathan/">LinkedIn. Designer drugs 2.0: AI ushers in a new era of drug discovery by Dr. Jonathan.</a></p><p>21. <a href="https://www.linkedin.com/pulse/accelerating-drug-discovery-how-ai-streamlining-process-sudharsan-b-l-bgqec/">LinkedIn. Accelerating drug discovery: How AI is streamlining the process by Sudharsan B.</a></p><p>22. <a href="https://datahorizzonresearch.com/artificial-intelligence-in-biotechnology-market-2333">DataHorizzon Research. Artificial intelligence in biotechnology market.</a></p><p>23. <a href="https://www.linkedin.com/pulse/artificial-intelligence-drug-discovery-development-recruitreeindia/">Recruitree India. Artificial intelligence in drug discovery and development. LinkedIn.</a></p><p>24. <a href="https://www.mckinsey.com/featured-insights/the-next-normal/biotech">McKinsey &amp; Company. Biotech and the next normal.</a></p><p>25. <a href="https://www.deloitte.com/uk/en/Industries/life-sciences-health-care/collections/life-sciences-and-health-care-predictions.html">Deloitte. Life sciences and health care predictions.</a></p><p>26. <a href="https://www.mckinsey.com/industries/life-sciences/our-insights/how-ai-could-revolutionize-drug-discovery">McKinsey &amp; Company. How AI could revolutionize drug discovery.</a></p><p>27. <a href="https://www.mckinsey.com/industries/life-sciences/our-insights/we-can-invent-new-biology-molly-gibson-on-the-power-of-ai">Gibson, M. We can invent new biology: Molly Gibson on the power of AI. McKinsey &amp; Company.</a></p><p>28. <a href="https://pmc.ncbi.nlm.nih.gov/articles/PMC8994527/">National Center for Biotechnology Information. (2022). AI applications in drug discovery. PMC.</a></p><p>29. <a href="https://pubmed.ncbi.nlm.nih.gov/39284102/">PubMed. (2023). Recent advancements in AI-driven drug discovery.</a></p><p>30. <a href="https://pubmed.ncbi.nlm.nih.gov/39396920/">PubMed. (2024). Innovations in biotechnology and drug discovery.</a></p><p>31. <a href="https://www.labiotech.eu/best-biotech/ai-drug-discovery-companies/">LaBiotech. Top AI drug discovery companies.</a></p><p>32. <a href="https://www.securities.io/ai-and-digital-biotech-companies/">Securities.io. AI and digital biotech companies.</a></p><p>33. <a href="https://www.owkin.com/](https://www.owkin.com/">Owkin. Owkin: Accelerating drug discovery with AI.</a></p><p>34. <a href="https://www.cbinsights.com/research/ai-readiness-index-pharma/">CB Insights. (2023). Pharma AI Readiness Index: Who’s best-positioned for the AI boom?</a></p><p>35. <a href="https://www.birlasoft.com/articles/demystifying-explainable-artificial-intelligence">Birlasoft. Demystifying Explainable Artificial Intelligence.</a></p><p>36. <a href="https://arxiv.org/abs/2303.00240">Arxiv.org. “Machine Learning Models for Predicting Binding Affinities of DrugBank Compounds to Opioid Receptors.” 2023. Accessed November 22, 2024.</a></p><p>37. <a href="https://bmcbiol.biomedcentral.com/articles/10.1186/s12915-024-02028-3">BMC Biology. “DrugReAlign Framework: Utilizing Large Language Models for Drug Repurposing.” 2024. Accessed November 22, 2024.</a></p><p>38. <a href="https://link.springer.com/article/10.1007/s00044-024-03253-9.">SpringerLink. “AI-Based Molecular Docking for Predicting Binding Affinities and Poses.” 2024. Accessed November 22, 2024.</a></p><p>39. <a href="https://hitconsultant.net/2024/11/13/nventures-invests-in-genesis-therapeutics-ai-platform-for-drug-development">Hit Consultant. (2024, November 13). NVentures invests in Genesis Therapeutics’ AI platform for drug development.</a></p><p>40. <a href="https://techcrunch.com/2021/06/22/a-i-drug-discovery-platform-insilico-medicine-announces-255-million-in-series-c-funding">TechCrunch. (2021, June 22). AI drug discovery platform Insilico Medicine announces $255 million in Series C funding.</a></p><p>41. <a href="https://www.mckinsey.com/industries/life-sciences/our-insights/generative-ai-in-the-pharmaceutical-industry-moving-from-hype-to-reality">McKinsey &amp; Company. (2024). Generative AI in the pharmaceutical industry: Moving from hype to reality.</a></p><p>42. <a href="https://www.reuters.com/technology/artificial-intelligence/nvidia-backed-ai-firm-iambic-unveils-drug-discovery-breakthrough-2024-10-29">Reuters. (2024, October 29). NVIDIA-backed AI firm Iambic unveils drug discovery breakthrough.</a></p><p>43. <a href="https://techcrunch.com/2024/04/24/xaira-an-ai-drug-discovery-startup-launches-with-a-massive-1b-says-its-ready-to-start-developing-drugs">TechCrunch. (2024, April 24). Xaira: An AI drug discovery startup launches with a massive $1B, says it’s ready to start developing drugs.</a></p><p>44. <a href="https://www.thetimes.co.uk/article/ai-new-drugs-and-the-appliance-of-life-science-p8q0sw8j0">The Times. (2024). AI, new drugs, and the appliance of life science.</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=05d2d9841ee1" width="1" height="1" alt=""><hr><p><a href="https://medium.com/included-vc/investing-in-biotechs-ai-gold-rush-what-smart-money-needs-to-know-about-the-future-of-medicine-05d2d9841ee1">Investing in Biotech’s AI Gold Rush: What Smart Money Needs to Know About the Future of Medicine</a> was originally published in <a href="https://medium.com/included-vc">Included VC</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
    </channel>
</rss>