<?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 Jessa Parette on Medium]]></title>
        <description><![CDATA[Stories by Jessa Parette on Medium]]></description>
        <link>https://medium.com/@jessa.parette?source=rss-1bd4a89f34e4------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*6XL7ikWzZDb0fMhwiTsBYQ.jpeg</url>
            <title>Stories by Jessa Parette on Medium</title>
            <link>https://medium.com/@jessa.parette?source=rss-1bd4a89f34e4------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Mon, 16 Nov 2026 06:54:11 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@jessa.parette/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[Why Voice AI Still Feels Wrong (Part 2): The Five Laws of Invisible Conversation]]></title>
            <link>https://medium.com/@jessa.parette/why-voice-ai-still-feels-wrong-part-2-the-five-laws-of-invisible-conversation-9e95a08130d9?source=rss-1bd4a89f34e4------2</link>
            <guid isPermaLink="false">https://medium.com/p/9e95a08130d9</guid>
            <category><![CDATA[voice-ai]]></category>
            <category><![CDATA[product-design]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[design-leadership]]></category>
            <category><![CDATA[ux-design]]></category>
            <dc:creator><![CDATA[Jessa Parette]]></dc:creator>
            <pubDate>Fri, 31 Jul 2026 08:17:04 GMT</pubDate>
            <atom:updated>2026-07-31T08:17:04.940Z</atom:updated>
            <content:encoded><![CDATA[<h3>The hidden principles of human conversation — and why every Voice AI team eventually runs into them.</h3><p>Think back to a conversation where you could hear, sense even, that something changed midway through. A sigh, an extra pause or a phrase spoken in a distracted voice, and suddenly you’re recalibrating, translating what is about to happen through the filter of what was just said.</p><p>Sensing the change in conversation through tone or inflection is an instinct practically stamped into our DNA.</p><p>Now, as AI reaches into one of our oldest, most human interfaces — the voice — nuances humans earned through millions of years in evolution have suddenly been bypassed for a technology that is taught to sound human.</p><p>Everyone is talking and writing about AI as if the technology has had a pedagogical status in every college curriculum for the last twenty years, yet the truth is: most of us have no idea what we are doing with it.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1008/0*USKXHwwPNAIbq-p0.jpg" /><figcaption>Jessa Parette presenting The Five Laws of Conversation at Design Leadership Summit 2025</figcaption></figure><p>In my<a href="https://medium.com/@jessa.parette/designing-for-porosity-the-missing-layer-in-voice-ai-c53016a93dc5"> last article</a>, I introduced the concept of porosity — the patterns of rhythm and sound that carry meaning beneath the words in any conversation. That gap, the one between complete mastery and novice, is exactly where <em>conversational porosity</em> lives, and exactly where voice AI tends to fail.</p><p>Of course, naming a problem is the easy part. What I propose here is not the ultimate solution, but a design principle approach for building voice AI systems that respect the most human aspects of conversation.</p><p>What follows is that framework: The Five Laws of Invisible Conversation.</p><ul><li><strong>Law I: The Emotion Principle</strong></li><li><strong>Law II: The Rhythm Law</strong></li><li><strong>Law III: The Tone Effect</strong></li><li><strong>Law IV: The Repair Rule</strong></li><li><strong>Law V: The Connection Constant</strong>.</li></ul><p>These laws are grounded in linguistic, psychological, and conversational science. They are not stylistic preferences. They are the structural conditions under which human conversation functions. A voice system that ignores them doesn’t produce a suboptimal experience. It produces one that feels, at a fundamental level, wrong.</p><p>I first presented these at the<a href="https://www.youtube.com/watch?v=wHEu-c7vTnQ"> Design Leadership Summit in 2025</a>. I’m writing them down here because disciplines are built in public — and this one needs to start somewhere.</p><h3>Law I: The Emotion Principle — Every sound carries feeling.</h3><p>During undergrad, a language professor asked me a question that has stayed with me.</p><p>“How many different meanings can you make from the sentence ‘<em>I didn’t take her wallet’</em> by changing a single inflection?”</p><p>Try it out, stress one word at a time, and you get five different meanings.</p><p>Five words. Five completely different meanings. None of them live in the vocabulary. All of them live in the prosody.</p><p>This is not a parlor trick. It is the entire problem with voice AI.</p><p>Human vocalization encodes structured emotional information beyond lexical content. UC Berkeley research <a href="https://bsil.studentorg.berkeley.edu/wp-content/uploads/2019/01/CowenEtAlAmPsych.pdf">mapped 24 distinct emotions</a> carried by <a href="https://news.berkeley.edu/2019/02/04/audio-map-of-exclamations">vocal bursts alone</a> — brief, non-verbal sounds that communicate meaning<a href="https://doi.org/10.1016/S0167-6393(02)00084-5"> across cultures</a> without a single word. When someone sighs, you know they’re frustrated. When they gasp, you know they’re surprised. No sentence required. The meaning arrives before the language does.</p><p>Voice AI systems that ignore affective signals are not interpreting conversation. They are transcribing it. Flat tone is not neutral. It is emotionally absent. And emotional absence degrades trust steadily, interaction by interaction, in ways that surface as churn long before anyone identifies the cause.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*QvUWgpodGjddAPFX.jpg" /><figcaption>Jessa Parette presenting The Five Laws of Conversation at Design Leadership Summit 2025</figcaption></figure><p>This means treating prosody — pitch, pace, intensity — as a design input from day one, not a layer added after the words are already decided. Add a prosody feature layer to your NLU to detect affect before choosing the next dialog act, especially in recovery scenarios. Design expressive output, not merely accurate output. The goal is a system that understands the moment, not just the words.</p><h3>Law II: The Rhythm Law</h3><p>I was listening to a usability recording of a woman ordering food from a voice AI system when she suddenly had to sneeze. Mid sentence, she pulled the phone away going, “…and I’ll haa-aaachoo”.</p><p>By the time she had put the phone back to her ear, a few seconds had passed. The voice AI system, most likely assuming the silence meant she was done, went ahead and answered: “Alright, I’ll place your regular order” — not the order she was building, the one already on file.</p><p>Any person on that call would have known exactly what just happened and waited for her to come back to “and I’ll have.” That’s not a hard read — a sneeze doesn’t sound like someone finishing a thought, and most who’ve ever had a phone conversation don’t need it explained to them.</p><p>The AI didn’t make that read. It heard silence where a turn used to be talking, and treated the silence as her handing the floor back. This is what happens when a system times conversation instead of listening to it as a human would.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/978/0*lxILtJMBdSaX0tFW.jpg" /><figcaption>Jessa Parette presenting The Five Laws of Conversation at Design Leadership Summit 2025</figcaption></figure><p>Why? It’s because of <a href="https://doi.org/10.2307/412243">a little phenomenon</a> in human conversation called “turn clusters”, where humans take ‘turns’ passing conversational rhythm back and forth. Like little batons, we pass rhythm back and forth through small silences, taking our cue of when to move forward by where the last person left off.</p><p>This turn-taking is so intrinsic it happens in a window of roughly 200 milliseconds, which is one fifth of a second. That speed — one fifth of a second — is faster than most conscious thought, making this ‘turn-taking’ a biological <a href="https://www.pnas.org/doi/10.1073/pnas.0903616106">wiring humans </a>naturally enact. We read it the same way we read a sneeze as a sneeze and not a sentence ending. Voice AI has access to the silence, but not what the silence means.</p><p>These rapid exchanges of short, rhythmically aligned sequences feel alive and connected. When that rhythm breaks — when there is a longer-than-expected silence — the human brain reads it as a social signal. Hesitation. Confusion. Disengagement. Even rejection.</p><p>A voice assistant that pauses two seconds before responding doesn’t feel slow, it feels broken. The gap that is imperceptible in a data pipeline is, in conversation, a relationship event.</p><p>This means latency is not a technical metric, but a design problem and a relational signal. Silence is social glue — until it isn’t. When response time is an engineering problem, the solution is optimization: Make it faster.</p><p>When response time is a design problem, the solution is intentionality. Defining what silence means, when it’s acceptable, what it communicates at the edge of the threshold. Moreover, defining when silence means ‘I’m thinking’ versus ‘I didn’t hear you’ or ‘I don’t care’.</p><p>Those are three different design decisions dressed up as the same technical spec. This distinction matters enormously for how teams prioritize.</p><p>To account for this, building voice AI means you must define silence thresholds as part of the UX specification, not only the engineering spec. Treat turn-taking as a rhythmic structure, not input/output. Enable interruption, overlap, and barge-in. Minimize system response delays to stay within the 200–250 millisecond window wherever possible.</p><p>And when you can’t stay within the window — design what the silence means. Don’t leave it blank.</p><h3>Law III: The Tone Effect — Tone predicts trust.</h3><p>Did you know that it takes less than 10 seconds for a human to predict someone’s trustworthiness based on tone alone?</p><p>In <a href="https://pubmed.ncbi.nlm.nih.gov/12110787/">a medical study</a>, researchers took <a href="https://www.sciencedirect.com/science/article/abs/pii/S0039606002000223">brief 10-second</a> clips of surgeons speaking, removed all lexical content so only tone remained, and found that surgeons with a harsh vocal tone had significantly more malpractice claims than those with a warm one.</p><p>Ten seconds. No words. Outcome predicted.</p><p>Now think about the voice AI your organization deployed last quarter. Think about the tone it uses when it says “I’m sorry, I didn’t understand that.” Think about whether anyone in the room when you shipped it asked: “Does this sound like it knows what it’s doing?”</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/710/0*iUVLpwSoKh0YCdW9.jpg" /><figcaption>Jessa Parette presenting The Five Laws of Conversation at Design Leadership Summit 2025</figcaption></figure><p>Prosodic features — pitch contour, warmth, pacing, sharpness — carry as much weight as the words themselves. Two identical sentences, “I can help with that,” delivered with different tones, can mean drastically different things and therefore lead to drastically different levels of trust. Humans analyze and respond to this shift in tone unconsciously, automatically, the way humans always have.</p><p>For organizations deploying voice AI at scale, this is not a design nuance. It is a brand risk, a liability surface, and a strategic variable — almost always unmanaged.</p><p>This implication means tone is a design system decision, not a voice talent one. Designing for this means operationalizing tone as a design-system primitive — standardized across contexts the way typography and color are standardized in visual systems. It requires defining contextual tone matrices, mapping emotional states to pacing and establishing explicit failure-tone specifications.</p><p>You wouldn’t ship a product where every team chose their own typeface based on personal preference. You shouldn’t ship voice AI where every flow has its own emotional register based on whoever wrote the script that week.</p><h3>Law IV: The Repair Rule</h3><p>My dad used to tell me, growing up bilingual, “Another language is another life.” I spent a lot of my childhood standing in the gap between two of them — translating, watching each person’s meaning arrive before the words meant to carry it did.</p><p>One afternoon I was translating for a family friend, Faiya, at a market stall. She pointed to a red purse. The vendor named a price. Faiya’s shoulders went up, just slightly, and she said, “Huh?”</p><p>No translation needed. The vendor understood her instantly — switched to broken English — tried the number again, adjusted. He heard the pitch of her voice go up on ‘Huh?’ and knew, before she’d said anything, there was a misunderstanding somewhere.</p><p>Human<a href="https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0078273"> dialogue anticipates misunderstanding</a>, and we <a href="https://doi.org/10.2307/413107">naturally</a> reach for the same type of ‘repair’ with a simple “Huh?”. Voice, therefore, comes before language.</p><p>This means emotion or communication starts as a sound, not as actual words understood. After all, babies come wailing into the world, screaming is their only form of communicating hunger, pain or fear. By the time the child is a few years older, they’ve hopefully replaced screaming with actual words like “Please change my diaper.”</p><p>The ambiguity is now gone. You know where that smell is coming from and what the kid needs.</p><p>Many voice systems respond to this same kind of ambiguity with: “Sorry, I didn’t get that.” That is regression, not recovery, halting the conversation rather than repairing it. It also treats misrecognition as exceptional when, in any high-volume deployment, misrecognition is statistical certainty. A system designed as though errors are exceptional will fail gracefully — exactly once — before users stop trusting it.</p><p>The difference is architectural and intentional. Apologizing is a script — something you write after you’ve decided errors are embarrassing. Navigating is a system — something you build before you launch because you’ve accepted that errors are inevitable and designed accordingly.</p><p>What this means is that voice AI builders need to create a resilience layer — confidence thresholds, clarification prompts, escalation pathways, modality shifts — designed in advance and governed cross-functionally across product, engineering, and operations.</p><p>Robust systems navigate the ambiguity that comes with being so very human, they don’t apologize for it.</p><h3>Law V: The Connection Constant</h3><p>Imagine you’ve had a hard day and meet a friend for drinks.</p><p><em>Scene one:</em> After venting, you pause. Your friend waits a second too long before saying, “Yeah, that sounds rough.”</p><p><em>Scene two:</em> Same scene, same friend, same issue. Except this time, when you pause your friend immediately responds with “Yeah, that sounds rough.”</p><p>Whether intentional or not, which version of your friend do you feel was in the moment with you? Connected to what you were saying?</p><p><strong>Law V, The Connection Constant states: “The smaller the silence, the bigger the connection.”</strong></p><p>This one is the simplest law. It is also the most ignored. Response latency correlates<a href="https://www.pnas.org/doi/10.1073/pnas.0903616106"> directly with felt connection</a> — the <a href="https://doi.org/10.1037/0022-3514.79.6.941">faster we respond</a>, the closer we feel. Even a small difference in timing changes how present someone feels on the other end of the line.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*vZsvk7KkIPPkhYMF.jpg" /><figcaption>Jessa Parette presenting The Five Laws of Conversation at Design Leadership Summit 2025</figcaption></figure><p>Silence is not just a pause. It is a signal. Too long, and the bond starts to fade. Immediate acknowledgment — a “Mm,” a “Right,” a brief backchannel — strengthens relational alignment in ways that content alone cannot replicate. When a conversational agent delays acknowledgment, users report lower perceived attentiveness regardless of how good the eventual response is. Connection is not solely semantic. It is rhythmic.</p><p>In voice AI, timing is the interface.</p><p>What does this mean for voice AI builders? Treat acknowledgment as interaction, not filler. Design backchannels — micro-responses that signal listening — as meaningful intents, not afterthoughts. Test pacing, silence, and warmth the same way you test task completion. Measure trust as a KPI. A system that only speaks when it has something complete to say is missing half of what conversation actually is.</p><h3>Why These Laws Operate as a System</h3><p>The five laws are not independent. They cascade.</p><p>Emotion undetected (Law I) produces a timing miscalibration (Law II). Miscalibrated rhythm distorts tone perception (Law III). Degraded tone makes repair feel inadequate (Law IV). Failed repair severs connection (Law V). The failure sequence is predictable — and so is the path back.</p><p>This is why designing for porosity is not a single design decision. It is a systems commitment. You cannot fix a voice AI that feels robotic by adjusting one variable. The invisible layer has to be designed holistically, or the cascade will find the weakest point.</p><p>When the screen disappears:</p><ul><li>Memory becomes the UI.</li><li>Timing becomes hierarchy.</li><li>Tone becomes brand.</li><li>Recovery becomes trust infrastructure.</li></ul><p>Without governing principles, conversational systems collapse into scripted automation — technically capable, relationally inert. With principles, they begin to approximate something more durable: fluency.</p><h3>The Real Design Challenge</h3><p>Voice is our oldest interface. Before we had words, we had tone. Before we had touchscreens, we had touch.</p><p>As design leaders, our legacy won’t be how beautiful our screens were. It will be whether the systems we built could understand us — whether we taught AI not just to respond, but to relate. Because every word, every pause, every sigh — that’s design too.</p><p>The Five Laws of Invisible Conversation are a first draft. They will be challenged, extended, and refined. That is the point. Disciplines are built in public, by practitioners who are willing to name things before the naming feels safe.</p><p>The organizations that codify these standards now will define how machines are allowed to sound in healthcare, commerce, retail, and public life for the next decade. That is not a narrow product decision. It is a cultural one.</p><p>Because when the interface disappears, all that’s left is <strong>us</strong>. Our tone, our timing, our <strong>humanity</strong>.</p><p><em>Jessa Parette is Head of Design at Byte by Yum! and an advisor to design and product leadership teams. She works at the intersection of AI-driven systems, organizational complexity, and experience design at scale.</em></p><p><em>This piece was originally published on </em><a href="https://www.jessaparette.com/articles/own-the-room-or-leave-it"><em>www.jessaparette.com</em></a><em>.</em></p><p><em>The Five Laws of Invisible Conversation were developed and first presented by Jessa Parette at the </em><a href="https://www.youtube.com/watch?v=wHEu-c7vTnQ"><em>Design Leadership Summit</em></a><em>, 2025. If you reference this framework, please cite it as: Parette, J. (2025). The Five Laws of Invisible Conversation: A Design Framework for Voice AI. Design Leadership Summit.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=9e95a08130d9" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The Drive-Thru is a Design Problem]]></title>
            <link>https://medium.com/@jessa.parette/the-drive-thru-is-a-design-problem-a3b4a16e3515?source=rss-1bd4a89f34e4------2</link>
            <guid isPermaLink="false">https://medium.com/p/a3b4a16e3515</guid>
            <category><![CDATA[service-design]]></category>
            <category><![CDATA[design-leadership]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[systems-thinking]]></category>
            <category><![CDATA[customer-experience]]></category>
            <dc:creator><![CDATA[Jessa Parette]]></dc:creator>
            <pubDate>Fri, 17 Jul 2026 07:36:33 GMT</pubDate>
            <atom:updated>2026-07-30T09:24:58.427Z</atom:updated>
            <content:encoded><![CDATA[<h4>Why faster technology hasn’t fixed the drive-thru — and why the most important system in the lane isn’t owned by anyone.</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*2zuaRDKVunPfdoz6.jpg" /></figure><p>I was standing in the back of house during a QSR site visit, headset on, listening to the drive-thru line work.</p><p>The AI was handling it cleanly. A customer pulled up to the speaker, said what they wanted — no fumbling, no repeating, no “I’m sorry, could you say that again?” The order moved through. The customer pulled forward. From where I was standing, it looked like exactly what the industry has been promising: frictionless, fast, almost elegant.</p><p>Then the POS went down.</p><p>What happened next had nothing to do with artificial intelligence. The manager on duty became something closer to a circus acrobat — physically running between the front counter and the drive-thru station, manually reconstructing each order, conducting ten things simultaneously with no baton and no score. The AI had done its job. The system around it had not.</p><p>I stood there, headset still warm from the conversation the technology handled perfectly, watching a human sprint to hold together what the technology couldn’t see.</p><p>Then it hit me: the drive-thru doesn’t have a voice AI problem. It has a systems design problem.</p><h3>The Four Systems</h3><p>The drive-thru is not one system. It is five — and the industry has spent fifty years optimizing four of them.</p><p>The physical lane. The communication chain that moves a spoken order into a kitchen. The production system that turns that order into food. The technology layer — POS, KDS, AI, forecasting — that ties the others together, until it doesn’t.</p><p>Every major QSR brand has people whose entire job is one of those four systems. Facilities owns the lane. IT owns the technology. Operations owns production. Vendors own the communication hardware. Decades of optimization, billions of dollars of investment, and a 25-year industry study tracking every second of service time.</p><p>The drive-thru accounts for an estimated<a href="https://www.qsrmagazine.com/story/the-2025-qsr-drive-thru-report/"> 50–70% of revenue</a> at major QSR chains — and <a href="https://www.prnewswire.com/news-releases/from-trend-to-transformation-off-premises-dining-now-essential-for-restaurant-consumers-operators-302429894.html">with 75% of all restaurant traffic</a> happening off premises, it is no longer a convenience feature. It is the dominant commerce format of American eating. And it is slowing down. Average service time has climbed to over four minutes — up from <a href="https://www.qsrmagazine.com/story/the-2025-qsr-drive-thru-report/">sub-three-minute benchmarks</a> a decade ago — not because operations have gotten worse, but because menus have gotten more complex.</p><p>The drive-thru ordering moment was designed around a specific set of assumptions — a driver, a static menu, a single-lane queue, a human on the other end of a speaker. None of those assumptions are universally true anymore. Mobile orders enter the lane mid-queue. Multi-passenger vehicles produce competing voices. Menus have grown from dozens of items to hundreds. And the speaker box — degraded, acoustically unengineered, unchanged in its fundamental interaction model for fifty years — is still the primary interface between a customer and a $60 billion commerce channel.</p><p>This is not a speed problem. It is a complexity problem that has been mistaken for a speed problem — which is why every technology intervention for the last decade has been aimed at throughput rather than experience. Faster AI. Faster screens. Faster timers. The lane gets instrumented. It does not get designed.</p><p>And still, the average drive-thru takes over four minutes.</p><p>The design question isn’t how to make AI work at the speaker. It’s what the ordering moment should actually be.</p><h3>What The Drive-Thru is Missing</h3><p>By the time a car has been sitting in line for three minutes without reaching the speaker, something has already shifted. I’ve watched it happen from the lane — the small ticks first, a head shake, fingers drumming, the phone coming out. Then one of two things: the car noses out entirely, or it stalls when the car ahead moves because the driver has already left mentally. The order isn’t lost at the speaker. It’s lost in the queue.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/766/0*_54I6zM0OWjdWj4O.jpg" /></figure><p>Science has a name for this phenomenon. Queue psychology research has documented for decades that <a href="https://pubsonline.informs.org/doi/10.1287/opre.35.6.895?__cf_chl_f_tk=Mbe1ABj65TuDJkA6hmoL0flrx7LI1aJ73FQFCvY1SsQ-1783414891-1.0.1.1-9YfCsF5fpVn5xdJH3G4hhhaScb_UMaA3iWHBuSrU9E0">perceived wait time inflates</a> with every minute past three — and that <a href="https://www.columbia.edu/~ww2040/4615S13/Psychology_of_Waiting_Lines.pdf">once the five-minute</a> mark is crossed, perceived duration drastically inflates. The drive-thru, with an average service time now exceeding 5 minutes, is operating inside that inflation zone for most of its customers, most of the time. It is not a coincidence that abandonment rates climb in the pre-speaker portion of the lane. The <a href="https://www.researchgate.net/profile/Richard-Larson-9/publication/304582002_Prescription_for_the_Waiting_in_Line_Blues_Entertain_Enlighten_Engage/links/5774627708aeb9427e2422c4/Prescription-for-the-Waiting-in-Line-Blues-Entertain-Enlighten-Engage.pdf?__cf_chl_f_tk=0wiuhAX.DSxEYW0GNN5IHdzfSa8eC6b_0O5JoLsZVKg-1783415031-1.0.1.1-g_nrRmdr9tFZrK2fdgUp5S.EH4H2FWI6ieh6Ll2gijI">science is specific</a>: pre-process waits — the time before service begins — feel measurably longer and generate more frustration than in-process waits. The minutes before the speaker are experienced as longer than the minutes after it. The queue itself is designed to feel worse than it needs to.</p><p>And then the customer reaches the speaker.</p><p>What happens at that moment is one of the most cognitively compressed interactions in everyday commerce. Time pressure. Social pressure — the car behind has been waiting too. Auditory interference from road noise, engine noise, a speaker system that has been dropped and degraded and never acoustically engineered for the environment it lives in. And a digital menu board that, in the worst cases I’ve seen from the customer side, presents a wall of promotional graphics where the actual menu used to be — because the marketing team pushed a campaign to the screen without anyone checking whether the base products were still visible. The result is a customer who pulls up ready to order and instead says: <em>“Uh… can I have the… where is your…”</em></p><p>If an interface makes something notably worse for the user, the fault is in the design, not the customer.</p><p>From inside the system, the failure compounds. I’ve seen digital menu boards running promotions that aren’t connected to the POS — the screen shows one thing, the register doesn’t know it exists, and a team member has to manually reconstruct a customized order in real time because nobody thought to push the same campaign to both systems. The customer experiences a delay they can’t explain. The team member absorbs a pressure spike the technology created. The gap between what the screen promised and what the system can deliver lands, invisibly, on a human being.</p><p>Voice AI carries the same pattern into the ordering moment itself. Voice AI can handle 72% of orders end-to-end — but achieves 81% accuracy doing it. When a human employee <a href="https://go.intouchinsight.com/hubfs/website/intouch-insight/resources/reports/drive-thru/drive-thru-study-report-2025.pdf">steps in after AI</a>, accuracy jumps to 95%. The technology is optimized for the happy path. The human judgment that catches what the AI missed is what makes the lane actually function.</p><p>The McDonald’s-IBM pilot made this visible at scale. Between 2021 and mid-2024, <a href="https://www.abacusnews.com/mcdonalds-archiq-google-ai-drive-thru/">McDonald’s deployed IBM’s Voice AI</a> to more than 100 U.S. drive-thru locations. The system failed publicly — nine sweet teas instead of one, random items added to orders, adjacent lane orders crossed. <a href="https://www.cnbc.com/2024/06/17/mcdonalds-to-end-ibm-ai-drive-thru-test.html">McDonald’s ended the pilot</a> in June 2024.</p><p>While sophisticated, algorithms tend to miss one crucial environmental reality: the drive-thru is an acoustical warzone. Engine rumble, wind pressure against the microphone, car radios bleeding, competing speech, passengers talking over each other — all this creates a cacophony of chaos for the technology. Research on <a href="https://dl.acm.org/doi/10.1016/0167-6393(94)00059-J">automatic speech recognition</a> has documented for decades that systems achieving over 95% accuracy in controlled environments can fall below 50% in real-world noise conditions.</p><p>A system designed for this environment requires advanced beamforming — microphone arrays that create a spatial focus on the driver’s voice, filtering everything else out. The IBM system processed every sound it could hear. That is how one car’s order ended up on another car’s tab.</p><p>The lane was never modeled as a design environment. The technology was built for a conversation. It was deployed into a construction site.</p><p>We have thirty years of science documenting what makes waiting tolerable and decisions easier, yet none of it seems to have permeated the average drive-thru’s design. Why? Not because of an an absence of operational expertise, but an absence of experiential design.</p><h3>The Accessibility Silence</h3><p>The drive-thru has always had a user it wasn’t designed for. It just never had to admit it.</p><p>I was at a BBQ chain on a Sunday — post-church rush, the busiest hour of the week. The owner, Jose (name changed), told me something that stopped me mid-conversation: “The older crowd hardly ever uses the drive-thru, and the ones that do — I end up repeating myself longer than it takes for them to actually order anything.”</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*gWKNnObGJYBLxiZa.jpg" /></figure><p>I looked out at the dining room. Most of the dine-in customers were older. Several had mobility aids. A few wore hearing devices. They had made a choice — not about food, but about which format of commerce they could actually access.</p><p>A survey by Inclusion Solutions found that 42% of deaf respondents <a href="https://www.qsrmagazine.com/news/new-drive-thru-deaf/">reported leaving a drive-thru</a> without ordering. The speaker box — already acoustically degraded, already engineered for nothing — is the only interface available. There is no alternative input. There is no fallback. The format simply excludes people who can’t use it and calls that a preference.</p><p>Voice AI alone does not solve this. Deployed without non-standard speech patterns, regional accents, or alternative input modalities in mind, it may widen the gap rather than close it.</p><p>The McDonald’s-IBM system that struggled with background noise was the same system being asked to interpret the speech of elderly customers, non-native English speakers, and anyone whose voice doesn’t match the training data. Accuracy rates that look acceptable in aggregate can mask systematic failure for specific populations.</p><p>The drive-thru has treated accessibility as a compliance problem — the minimum required by law, applied at the minimum possible scale. That is a different design intent than asking: who can’t use this, and what would it take to include them? The distinction determines whether you design the floor or design for the full range of human variation.</p><h3>The Fifth System</h3><p>Because there is a fifth system. And nobody owns it.</p><p>It has no interface. No documentation. No dashboard. It runs entirely on human judgment — the continuous, real-time, invisible work of adapting the operation to whatever reality has decided to do today. Who bumps a ticket when two orders collide? Who pulls a car forward when production falls behind? Who reallocates labor at 12:07 on a Tuesday when three things fail at once?</p><p>The industry built technology on top of an assumption: that human judgement would handle everything the technology couldn’t anticipate.</p><p>And that assumption, that the gaps between systems would be handled by someone, means the technology works exactly as designed: with painful fissures that force operators to become the human integration layer.</p><p>Veteran operators think about this layer constantly. They model the failure path — because that’s where they live. The rest of the industry models the happy path. And then it deploys technology into it.</p><h3>What Systems Design Would Actually Look Like</h3><p>I want to tell you about a spatula.</p><p>I was in the back of house of a U.S. QSR location, doing what I always do in stores — watching how the system actually runs, not how it’s supposed to run. There was a spatula sitting on the counter near a food station. I picked it up to put it away. The manager stopped me.</p><p>She needed it.</p><p>I watched her take the spatula and use it to press the bump button on the kitchen display system — the screen mounted above the station that sequences orders through the kitchen. The button was too high for her to reach with her hand. So the spatula became the interface. Not a workaround, not a complaint — just how that kitchen runs. Nobody had filed a ticket. Nobody had flagged it. The KDS wasn’t broken. It was working exactly as designed. The design had just never accounted for the full range of people who would use it.</p><p>That spatula is the fifth system made physical. A gap filled by human ingenuity so quietly that it becomes invisible. And invisible gaps don’t get fixed — they get inherited.</p><p>The drive-thru has never had a systems designer in the room. It has had operations people, technology vendors, facilities architects, and marketing teams — each optimizing their own layer, handing off to the next, assuming someone else owns the boundary. Nobody does. The result is a $60 billion commerce channel held together by spatulas, sprinting managers, and the decision-making of whoever happens to be on shift.</p><p>Systems design would start by treating the four-minute drive-thru experience as a narrative, not a transaction. What should be true at the pre-speaker moment — when queue psychology says the customer is already losing patience? What should happen at the order moment — when cognitive load is highest and the interface is worst? What is the handoff designed to feel like, not just function like?</p><p>It would apply environmental design to a lane that has never had any. Acoustic engineering for a speaker system that has been degraded and ignored for decades. Menu board design that reduces cognitive load at the exact moment decisions have to be made. Signage that gives customers a sense of progress before they reach the speaker — the single cheapest intervention the research supports, and the one nobody has made.</p><p>It would design for the full range of people who use the lane, not the median. Not as a compliance exercise. As a design standard.</p><p>And it would finally design the human-AI handoff — the moment that may matter more than anything else in the lane right now, and the moment that currently receives the least design attention of anything in it.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/700/0*k-XXL6qv-fu-nx5H.jpg" /></figure><p>Every Voice AI deployment has an escalation path: the moment the AI can’t handle the interaction and a human takes over. In about <a href="https://www.intouchinsight.com/resources/studies/drive-thru/">21% of AI-assisted orders</a>, employees stepped in — most often because the AI couldn’t handle customization, answer a question, or process an unavailable item. That moment is currently an afterthought — a fallback, not a feature. It has no designed trigger, no designed tone, no designed experience for the customer who just watched a machine fail to understand them. It is the most sensitive moment in the AI-assisted drive-thru, and it has been left entirely to chance.</p><p>Designing that handoff — when it happens, how it’s communicated, who owns it, what the customer experiences in the transition — may be more important than optimizing AI accuracy by another two percentage points. Because the handoff is where trust is either recovered or lost. And right now, nobody owns it.</p><p>The brands that figure this out will not be the ones with the best AI. They will be the ones with the best designers — who finally showed up for the lane.</p><h3>At 2:40 PM</h3><p>At 2:40 pm, in the same location where I’d watched a manager sprint between the front counter and the drive-thru station hours earlier, her phone alarm went off. The lane was empty. Every customer in the dining room had already ordered. She walked to the kitchen and started a drop.</p><p>I asked her what she was prepping for.</p><p>“The high school lets out soon,” she said.</p><p>It took me a moment. She was dropping fried chicken at 2:40 so it would be fresh for the hold time — because she knew, from years of standing in that lane, that students would come through in about twenty minutes. No dashboard told her that. No forecast model flagged it. No AI prompted the drop. She just knew.</p><p>At 3:11 pm, two cars driven by teenagers came through the drive-thru. They knew her name. She asked about their grades.</p><p>Her judgement — dropping the chicken at 2:40 pm — was the highest-value labor decision made in that location all day. It is the fifth system at work, honed over years of building two-minute relationships through a speaker box. It is a system that doesn’t live in any database, yet is the most valuable thing operating in that lane.</p><p>The technology that keeps arriving to fix the drive-thru isn’t built by people who have worn the headset when the POS goes down, watched a manager become the integration layer between systems that were never designed to talk to each other, or seen a spatula forced to become the KDS interface.</p><p>They will keep building building systems that work in the demo and break in the lane.</p><p>Meanwhile, the fifth system will keep running as it always has: through un-captured operator intelligence that is acting as a human integration layer.</p><p>It deserves, finally, to be designed for.</p><p>_____</p><p><a href="http://www.jessaparette.com/"><em>Jessa Parette</em></a><em> is Head of Design at </em><a href="https://www.yum.com/wps/portal/yumbrands/Yumbrands/bytebyyum"><em>Byte by Yum!</em></a><em> and an advisor to design and product leadership teams. She works at the intersection of AI-driven systems, organizational complexity, and experience design at scale.</em></p><p><em>Originally published at </em><a href="https://www.jessaparette.com/articles/the-drive-thru-is-a-design-problem"><em>https://www.jessaparette.com</em></a><em> on July 17, 2026.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=a3b4a16e3515" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Own the room, or leave it]]></title>
            <link>https://uxdesign.cc/own-the-room-or-leave-it-39436e89a256?source=rss-1bd4a89f34e4------2</link>
            <guid isPermaLink="false">https://medium.com/p/39436e89a256</guid>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[design-leadership]]></category>
            <category><![CDATA[ux]]></category>
            <category><![CDATA[product-design]]></category>
            <category><![CDATA[career-advice]]></category>
            <dc:creator><![CDATA[Jessa Parette]]></dc:creator>
            <pubDate>Tue, 30 Jun 2026 09:27:23 GMT</pubDate>
            <atom:updated>2026-08-04T07:33:44.049Z</atom:updated>
            <content:encoded><![CDATA[<h4>Unlearning the habit that quietly dissolves design’s authority</h4><figure><img alt="A glass of water with dissolving black ink" src="https://cdn-images-1.medium.com/max/1024/0*GSc_6N16nvKIDeQV" /><figcaption>Photo by <a href="https://unsplash.com/@jeanpubi?utm_source=medium&amp;utm_medium=referral">Jingyu Liu</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><p>Once upon a time, I was a young designer in a meeting that could have been an email.</p><p>The problem was straightforward: a support team covering 30,000+ knowledge base articles across 14 countries needed a better way for people to find answers to common problems. Employees couldn’t find answers fast enough — losing time, making mistakes, escalating tickets that should have closed in two minutes. The proposal was clean, even elegant. A guided help feature layered on top of the existing knowledge base. Better search. Better surfacing. Faster answers.</p><p>This is when the meeting stopped being an email and became a war room.</p><p>Because I could see the problem. The knowledge articles were self-authored. No standards, no review process, some months out of date, some contradicting each other. A few were just wrong. The real fix was an audit — a slow, tedious, unglamorous crawl through every article before a single line of the new feature got built. And nobody in that room wanted to hear that. The timeline was set. The solution was decided.</p><p>My job, as far as the product leader was concerned, was to make it look good.</p><p>I was barely invited to that room in the first place. So when the product leader pointed to usage data and like/dislike counts as a way to weed out the unhelpful articles over time, I looked at it and thought: “<em>Ok. Good enough.”</em></p><p>The feature shipped. Support volume went up — by a lot.</p><p>Here’s what I’ve had to sit with: I didn’t fail to see the problem. I saw it clearly. I made a calculation, decided the partial solution was close enough, and let it go. That’s a different kind of mistake. Harder to name. Harder to defend. Because “good enough” is always a choice — and at a certain point in your career, it’s the most expensive one you can make.</p><p>Sound familiar?</p><p>The hard part isn’t recognizing the story. It’s recognizing that what felt like strategy was a habit. A well-dressed one. The kind that convinces you you’re reading the room when you’re actually just repeating a pattern you learned so early it stopped feeling like a choice.</p><p>That’s the part nobody names. And unnamed habits don’t get examined, they just get more expensive.</p><p>What took me years to understand: this wasn’t a problem title could solve. It cropped up again and again — different companies, different teams, different products. Studio designers, agencies, in-house or even freelancers. The context doesn’t matter. The problem wasn’t one of seniority, politics, or “we just need a seat at the table” — thought God knows we said that enough.</p><p>The issue wasn’t always just getting into the room, but also using design to <a href="https://medium.com/design-bootcamp/elevating-design-leadership-finding-the-right-flight-level-bc877d013bdd">own the altitude</a> of the problem.</p><p><strong>Altitude ownership is the refusal to let your discipline become optional</strong> — by naming what you own, defending what the work requires, and holding the stake even when the room would rather you didn’t.</p><p>The inverse of altitude ownership is abdication, because silence in a room where you hold influence isn’t neutral — there is a cost.</p><h3>Knowing the cost means knowing where you stand</h3><p>Design careers have altitudes. But altitude isn’t a ladder.</p><p>You don’t have to climb. Plenty of exceptional designers spend their entire careers at one level and do extraordinary work there. The principal designer who has spent fifteen years perfecting interaction craft at the interface level is not someone who failed to climb. They’re someone who chose mastery over altitude. And that choice produces something irreplaceable — the foundation that everything else gets built on.</p><p>The difference between levels isn’t seniority, salary or title. It’s one thing<em>: the ability to produce quality decisions under increasing uncertainty.</em></p><p>With each altitude, the uncertainty doubles. And the cost of being wrong scales with it. At the interface level, a bad call means you or your team has a bad day. At the product level, it means a feature fails and a roadmap shifts. At the systems level, it means something breaks that nobody saw coming — and people are scrambling to find out why. At the organization level, the wrong design bet doesn’t just cost a sprint. It can cost people their jobs. Their livelihoods. And you have to make that call anyway, with incomplete information, under pressure, in real time.</p><p>That’s what altitude actually means. Not where you sit in the org chart. How much uncertainty you can hold — and still make the call.</p><p><strong>The manual ends here</strong></p><p>At the beginning of most careers, there are manuals for everything. Any designer worth their pixels has probably wondered if they actually knew what they were doing or were just really good at searching Google. Say hello to imposter syndrome.</p><p>Heuristics. Accessibility standards. Design systems. Best practices. Entire libraries dedicated to helping you make good decisions. These are at the tips of our fingers when starting out, and most designers assume mastery means learning all of them</p><p>It doesn’t. Mastery is what happens when the manuals stop providing answers.</p><p>At some point, if you’re good enough, you’ll find yourself staring at a problem no manual can solve, no search can answer, and no expert seems particularly interested in solving for you. In that moment, the question isn’t whether you know the answer — it’s whether you’re willing to make a decision without one. This is the fork in the road that nobody talks about.</p><p>Some go deeper, becoming exceptional practitioners who master a discipline so completely the rest of us study their work. Others decide to operate where the manuals end, making decisions before certainty arrives — building the patterns and frameworks everyone else eventually adopts. Neither path is wrong. But they are fundamentally different altitudes of the same work.</p><p><strong>The judgement gap</strong></p><p>When I say altitude, I’m not talking about title, authority, or job level. I’m talking about scope — the combination <a href="https://hbr.org/2026/02/to-lead-through-uncertainty-unlearn-your-assumptions">of responsibility and uncertainty</a>. The higher the altitude, the fewer correct answers exist. Most assume that intelligence, expertise, or experience alone are what help people succeed at these higher altitudes — and yes, intelligence helps, and yes, you want a boss who actually has more experience. But those things alone aren’t sufficient. What separates people who make altitude transitions successfully from those who don’t is judgment. Not judgment as in being judgmental — judgment as in knowing what to do when the answer is incomplete, the information is conflicting, and the consequences are real.</p><p>Most designers assume advancement comes from acquiring new skills. Learn research. Learn facilitation. Learn business strategy. Learn AI. Learn whatever LinkedIn is excited about this week. Skills matter — but after watching designers successfully make altitude transitions and fail to make them, I’ve become convinced that skills are rarely the thing holding people back. It’s judgment. And this isn’t unique to design. Michael Watkins argues in “The First 90 Days” that one of the most common reasons leaders fail at transitions is because they continue doing the work that made them successful at the previous level. They aren’t failing because they’re incapable. They’re failing because the job changed and they didn’t.</p><p>Noel Tichy and Warren Bennis take the argument a step further. In <em>Judgment: How Winning Leaders Make Great Calls</em>, they argue that what separates effective leaders from ineffective ones isn’t intelligence or experience — it’s the ability to make quality decisions when the answer isn’t obvious.</p><p>Designers do this too. The interface designer keeps optimizing screens after becoming responsible for outcomes. The product designer keeps optimizing outcomes after becoming responsible for systems. The systems thinker keeps optimizing systems after becoming responsible for organizational risk. They’re solving yesterday’s problem at today’s altitude — because every altitude transition asks the same thing: <em>what do you do when no one knows the answer?</em></p><h3>Four Altitudes in a Design Career</h3><p>The answer is different at every altitude. In design careers, I’ve found there are largely four, each commanding their own scope of accountability and ownership: interface, product, systems and organizational.</p><figure><img alt="Copyright Jessa Parette 2026 — image of two axis, with uncertainty in the horizontal like and altitude names in the vertical" src="https://cdn-images-1.medium.com/max/1024/1*1E2tMuFS6q47DI1MyO5epw.jpeg" /><figcaption>Copyright © Jessa Parette · www.jessaparette.com</figcaption></figure><p>At the interface level, someone has to own the experience details, and that someone is design. Design owns the experience details not because it was assigned, but because it’s what the discipline requires. The moment a tradeoff forces those details to get cut, your job isn’t to accept the cut quietly. It’s to name who is now accountable. Who owns the corner that just got cut. Because if you don’t name it, you’ve answered the question by default — and the answer is nobody. When nobody owns it, design didn’t lose the argument. Design forfeited its stake entirely.</p><p>What doesn’t always exist is the authority to enforce it. A junior designer can flag a contrast ratio that fails accessibility standards and still watch an engineer ship it anyway, because nobody in the room outranks the deadline. That’s not a failure of ownership, because <strong>ownership and authority are two different things</strong>.</p><p>You can own the standard intellectually — it’s already written down, already documented, already true before you walked in the room. What you can’t always do is stop the cut. <strong>What you can always do is say it out loud.</strong> Name what’s being thrown away and who signed off on throwing it away.</p><p>Ownership at this altitude isn’t about having the final call. It’s about knowing what the work requires and refusing to let the decision get made silently.</p><p>This is also where the habit forms — see the problem, document it, wait for a better moment. This is not because interface-level judgement is weaker, but because it’s often the first place a designer learns that naming what the work requires can cost you the room. Amy Edmondson’s research on psychological safety helps explain why — environments that punish the wrong kind of friction train people to stay quiet, and that training doesn’t disappear when the title changes.</p><p>Design doesn’t lose ownership at this altitude because the work is easy. It surrenders it, one quiet deferral at a time — and that can happen to a principal designer with fifteen years of craft just as easily as someone six months in.</p><p><strong>Product altitude: Knowing how to own the outcomes</strong></p><p>The leap from interface to product is the leap from owning details to owning outcomes. The work stops being “is this screen right” and starts being “did this bet pay off.” And here is where most designers hit the same wall: the moment you try to define what “paid off” means, someone tells you that’s not your job.</p><p>It <em>is</em> your job. Just not in the way most designers think.</p><p>I once watched a designer run usability testing on a signup flow. The results were good — over 80% task completion. I asked one question: did they know what they were signing up for? Did you ask them to find a button, or find a solution? The designer stalled. The test measured whether people could complete a task. Not whether they understood what they were committing to. Those are not the same outcome, and nobody had decided which one mattered before the test was built.</p><p>I saw the same gap at a much larger scale years earlier, leading design at Walmart. There was a program that let employees bundle several existing perks if they opted to use their personal phone instead of a company-issued one. Most employees were already eligible. They just didn’t know it, because the perks were run by three separate teams that had never been stitched together. The product team wanted to measure how many people chose to bring their own device — that was the number that mattered to the bottom line. But that metric only captured a fraction of the actual problem. The real question wasn’t how many people switched devices. It was how many eligible employees ever found out they qualified for something they already had access to. I pushed to measure conversion into the bundle, not device adoption — because device adoption was the side effect, and the unclaimed benefit was the actual failure.</p><p>Neither room handed me the authority to define success. Someone had already decided what the number was supposed to be before I said anything.</p><p>The power wasn’t in overriding the metric. It was asking what the metric was a proxy for — because every number in that room was already a proxy for something nobody had bothered to define out loud. Roger Martin calls this “<a href="https://rogerlmartin.com/lets-read/the-opposable-mind">integrative thinking</a>” — the capacity to hold two opposing ideas simultaneously rather than defaulting to the easier one. In this case, the easier idea was the metric already on the table. Once you voice what the real number is hiding, the room can’t pretend it didn’t hear you.</p><p>That’s the unnamed power at this altitude. You may not own the final call on what gets measured. But you almost always have the standing to ask whether the metric on the table is measuring the real outcome or a comfortable proxy for it. Most designers never ask. They assume that’s a product question, not a design one. It’s both.</p><p><strong>Systems altitude: Making the invisible cost visible</strong></p><p>Design changes rarely live in isolation. There’s a cascading effect hidden in your working file.</p><p>A few years ago, before AI became what it is today, chatbots were the shiniest silver bullet for company problems. Long service desk wait times? Call center traffic jammed with the same question from 1,000 different users?</p><p>“Just throw a chatbot on it!” Is something I literally heard a VP in engineering say in a meeting.</p><p>But good designers know that designing a chatbot isn’t a surface change, it is a workflow triggering event. Adding a chatbot changes what gets logged, who reviews what, how escalations flow and what needs legal review if it fails “safe guardrails”.</p><p>None of that shows up in a design file. All of it shows up eventually, usually as someone else’s emergency, three months after launch.</p><p>The power here is <em>making the invisible cost visible</em> before it becomes someone else’s emergency. That sounds simple. It rarely is — naming a dependency means<a href="https://www.scientificamerican.com/article/kahneman-excerpt-thinking-fast-and-slow/"> naming it to someone who doesn’t report to you</a>, in a system you don’t control, before there’s any proof you’re right.</p><p>It gets harder once you have actual authority to act on what you see — because then the temptation isn’t silence. It’s speed.</p><figure><img alt="Copyright Jessa Parette 2026 — image of two axis, with cost of being wrong in the horizontal like and amount of uncertainty in the vertical" src="https://cdn-images-1.medium.com/max/1024/1*onAVFNYnSsgUCOj_LumaSA.jpeg" /></figure><p>I once owned the vendor relationship for a usability testing platform inside a heavily regulated financial institution. Designers on my team would ask me to approve plugins — small things, the kind that would have made their week meaningfully easier. I could approve them. Nobody was stopping me.</p><p>But greenlighting a plugin without first routing it through cyber and risk review wasn’t a shortcut. It was a decision to absorb liability that wasn’t mine to absorb alone, because a faster recruiting workflow for a usability study could just as easily become an unreviewed data practice that violates a “no contact” privacy law — and that’s not a design problem anymore. That’s an SEC-level problem wearing a design request as a disguise.</p><p>At each alititude, the most expensive design problems rarely look like design problems.</p><h3>Organizational altitude: where uncertainty becomes the work</h3><p>In a more recent meeting — one that could not have been an email — I got asked the question design leaders dream of and dread in equal measure.</p><p>“How will this research change what we planned for our Q2 roadmap for X product?”</p><p>That is not a design question. That is a capital allocation question and political landmine masquerading as a design question. Executives with honest answers come prepared with math, not design.</p><p>My honest answer in that meeting meant showing the cost of hiring a contractor versus reallocating a staff designer to the project, all at the assumption of what revenue the feature could unlock.</p><p>At an organizational altitude,<a href="https://uxmag.com/articles/head-of-design-is-dead-long-live-the-head-of-design"> design leaders are making very few design decisions</a>. The decisions become capital allocation, talent bets, org restructuring, board narratives — and design is the lens you bring, not the subject.</p><p>You’re not arguing for design anymore. You’re pricing it — translating a research finding or a craft instinct into the language the rest of the room actually allocates against: cost, time, population served, return.</p><p>The unnamed power at organizational altitude isn’t vision. It isn’t even risk tolerance. It’s <a href="https://hbr.org/2020/05/begin-with-trust">the discipline of making a call you can’t personally verify</a> — and building enough trust that the people who can verify pieces of it will move on your word <a href="https://www.scientificamerican.com/article/kahneman-excerpt-thinking-fast-and-slow/">instead of waiting for proof</a>.</p><h3>The cost of waiting</h3><p>There are times when waiting<em> is</em> the right call. Trust takes time to build. Credibility compounds with slow purposes. Relationships take time before your word alone becomes evidence enough to decide.</p><p>That is not the type of waiting I’m describing.</p><p>I’m describing hesitation that abdicates the one thing design is actually for — figuring out how something works, not just how it looks.</p><p>In the knowledge base story, waiting caused service tickets to spike. That was probably the cheapest consequence available at that altitude. Moving higher means the cost accumulates exponentially.</p><p>At the interface level, it costs a sprint, maybe two — something gets fixed and the team moves on. At the systems level, when naming the dependency before it’s on fire is key, hesitation might mean a legal review was missed or a key workflow wasn’t redesigned in time. Operating at the organizational level means hesitation might cost a market window or a competitive edge because your competitor built the thing you were still getting permission to study.</p><p>Decisions not made are still decisions. Accountability still lands somewhere — the only question is whether it lands on you now when you speak up about a tradeoff, or in three years when you scramble to explain why nothing happened.</p><p>The moment you see the problem — not the moment you’re given permission to speak — is the moment your leadership begins. That is the moment you have a choice. What you do next is the answer to every altitude question this piece has asked.</p><p><em>This piece was originally published on </em><a href="https://www.jessaparette.com/articles/own-the-room-or-leave-it"><em>www.jessaparette.com</em></a><em>.</em></p><p><em>Jessa Parette is Head of Design at Byte by Yum! and an advisor to design and product leadership teams. She works at the intersection of AI-driven systems, organizational complexity, and experience design at scale.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=39436e89a256" width="1" height="1" alt=""><hr><p><a href="https://uxdesign.cc/own-the-room-or-leave-it-39436e89a256">Own the room, or leave it</a> was originally published in <a href="https://uxdesign.cc/">UX Collective</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[AI just called design’s bluff]]></title>
            <link>https://uxdesign.cc/ai-just-called-designs-bluff-2b94bf2f5cae?source=rss-1bd4a89f34e4------2</link>
            <guid isPermaLink="false">https://medium.com/p/2b94bf2f5cae</guid>
            <category><![CDATA[ai-design]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[ux]]></category>
            <category><![CDATA[design-leadership]]></category>
            <category><![CDATA[product-design]]></category>
            <dc:creator><![CDATA[Jessa Parette]]></dc:creator>
            <pubDate>Tue, 12 May 2026 07:16:32 GMT</pubDate>
            <atom:updated>2026-07-17T11:13:57.574Z</atom:updated>
            <content:encoded><![CDATA[<h4>We spent a decade hiding behind the interface. Now it’s gone.</h4><p>Eight months ago, the questions started showing up in meetings I wasn’t supposed to notice them in.</p><p><em>“Can’t Figma AI just do that?”</em></p><p><em>“We used Claude for the mockups — do we still need to bring in a designer?”</em></p><p><em>“If the AI can generate the screens, what exactly are we hiring for?”</em></p><p>I am Head of Design at a global restaurant technology platform. I built design teams at Walmart and Capital One. I have spent years in rooms where I had to fight for design to be treated as something other than a screen factory — the team you call when engineers need something to build and the team you cut when the budget gets tight.</p><p>And for eight months, I have been fielding versions of the same question.</p><p>It stings. But here is the thing nobody wants to say out loud:</p><p>They are not entirely wrong to ask.</p><p>Not because design doesn’t matter. Because for most of the last decade, a significant portion of what design teams actually produced — the thing designers were hired, evaluated, and promoted for — was exactly the kind of work AI can now do faster, cheaper, and without a Slack message.</p><p>We built the conditions for this. We just didn’t think we were building a trap.</p><p><a href="https://www.petermerholz.com/newsletter/tma-how-design-succeeded-its-way-into-irrelevance/">Peter Merholtz recently wrote </a>what a lot of us in design leadership have been thinking but not saying out loud: design didn’t lose its strategic seat because organizations took it away. Design traded it — incrementally, for salary bands and delivery velocity and the comfortable feeling of being needed, right up until the moment it wasn’t. <strong>He’s right.</strong> I’ve been a design executive for over a decade at Walmart, Capital One, and now Byte by Yum!, and I have watched design leaders — myself included — make that trade without fully examining what we were giving up.</p><p>This is not a piece about Peter being wrong. It’s a piece about what comes next.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*60luMDrt-06DqeMv" /><figcaption>Photo by <a href="https://unsplash.com/@fertroulik?utm_source=medium&amp;utm_medium=referral">Fer Troulik</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><h3><strong>We Built This Trainwreck Ourselves</strong></h3><p>Design built the conditions of its own displacement starting in the 2010s when design thinking began its big boom. By the time I joined my first corporate job, I was inheriting the wreckage of that boom — corporate halls that saw design thinking as a workshop instead of a way to shape industry.</p><p>Design scaled. Teams of ten became teams of hundreds. Salaries doubled. Org charts moved design closer to the C-suite and it looked, from the outside, like design had finally arrived. We had seats at tables we had spent years being excluded from.</p><p>What was actually happening was narrower than that.</p><p>Design was scaling to support agile delivery. Every product team needed a designer. Every designer needed to keep engineers unblocked. The greatest professional sin was being the bottleneck — the person who slowed the sprint, the one whose work wasn’t ready, the designer who wanted to do more research before the thing shipped.</p><p>So we got fast. Really fast. We got good at producing screens quickly, consistently, in volume. We built design systems so nobody had to make decisions twice. We shipped. We hit our velocity numbers. We got promoted for it.</p><p>And while all of that was happening, something else was quietly leaving the building.</p><p>Not design. Design was everywhere. I mean the part of design that cannot be templated: the judgment that looks at a brief before a line of code is written and says <em>this will fail someone, and here is how</em>. The instinct that holds a whole system in your head while everyone else is staring at the feature in front of them. The willingness to say the uncomfortable thing in the room where the decision is actually being made.</p><p>Those capabilities did not disappear because organizations stopped valuing them.</p><p>They atrophied because we stopped practicing them. We were too busy shipping to build the muscles that would have made shipping matter.</p><p>I was in those rooms. I made versions of that trade. I know exactly what it felt like to choose velocity over the harder conversation — and to tell myself I would get back to the harder work later, when things settled down.</p><p>Things did not settle down. They never do.</p><p>Design didn’t get disrupted. It got what it built.</p><h3><strong>AI automates what is lowest value first. For design, that was UI.</strong></h3><p>Not because UI is easy. It isn’t. But because it is repeatable. Learnable. Describable enough to train a model on. The things that made a designer valuable inside an agile team — the ability to produce good enough material quickly enough, consistently enough, without dropping the sprint — are exactly the things that translate cleanly into a prompt.</p><p>The problem is that a significant number of designers built their entire professional identity around being fast at that.</p><p>Not because they were lazy. Because they were rewarded for it. Velocity was the metric. Screens shipped was the proof of value. The designer who could turn around ten high-fidelity mockups by Thursday was the one who got promoted, got the raise, got pointed to as the example. We built a profession that selected for throughput and then acted surprised when throughput got automated.</p><p>I watched it happen from inside it. I have sat in performance reviews where the conversation was entirely about delivery. How many products did the team touch. How quickly did designs move to handoff. Whether the design system was being used consistently enough to speed up the next sprint. Rarely did anyone ask: did we design the right thing? Did we see what was coming before it shipped? Did we prevent anything?</p><p>We did not have language for prevention. We had language for production.</p><p>So when AI arrived and production got cheap, a lot of designers did the only thing they knew how to do: they got faster. They learned the new tools. They optimized their prompts. They competed on the dimension that was being commodified.</p><p>That is the wrong bet.</p><h3><strong>AI didn’t expose a weakness in design. It exposed a lie we told ourselves to survive.</strong></h3><p>For thirty years, the field has said the same thing in every executive briefing, every conference keynote, every hiring conversation where we were trying to justify the budget: <em>design isn’t how something looks. It’s how something works.</em></p><p>We meant it. And we also couldn’t fully prove it, because the interface was always in the way.</p><p>The screen was the only thing that made the invisible visible. You couldn’t show a CEO how a system worked without first showing them what it looked like. You couldn’t get engineers aligned on a concept without mocking it up. You couldn’t get stakeholders to feel the experience without rendering it. So we rendered it. Constantly. We spent the majority of our professional hours producing the visual proof of thinking that lived somewhere deeper — because without the visual, nobody could see the thinking at all.</p><p>AI just made that part cheap.</p><p>And what’s becoming clear, now that the scaffolding is gone, is that for a lot of people who called themselves designers — the scaffolding was the only skill the job ever asked them to develop.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*rzJA71HaDxgUfeYj" /><figcaption>Photo by <a href="https://unsplash.com/@tzak94?utm_source=medium&amp;utm_medium=referral">Taylor Fry</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><p>I want to be precise here, because this is not an argument that UI is easy or that the people who do it well are somehow less than. UI is hard. Doing it well requires a depth of craft that takes years to build — visual hierarchy, interaction logic, accessibility, the thousand small decisions that make something feel effortless to use. That work matters. It has always mattered. But UI is the translation layer, not the thing itself.</p><p>What the last decade did, in its sprint toward velocity, was quietly stop rewarding everything else. The job became: be fast at the primary thing. Ship. Repeat. The T-shape that design education preached got flattened into a single vertical line.</p><p>But if you spent the last decade learning to show the work without learning to do the work — if the mockup was the thinking and not the artifact of the thinking — then what AI took from you wasn’t a tool.</p><p>It was the thing you were hiding behind.</p><p>I don’t say that to be cruel. I say it because I have sat across from talented, hardworking designers who are genuinely terrified right now, and the terror is specific. It is not <em>my tools are changing.</em> It is <em>I don’t know what I am without the tools.</em></p><p>That is the question the field needs to answer.</p><h3><strong>The market is repricing the value of design.</strong></h3><p>This is not a story about AI eliminating design. It is a story about companies being forced to change what they optimize for — and design needing to change with it. Right now, in most organizations, the optimization is around speed, risk reduction, margin, and adaptability. The ability to move fast without breaking things that cost a fortune to fix later.</p><p>In that environment, certain work gets automated first. Not because it’s unimportant — but because it’s describable. Repeatable. Trainable. Production work. Visual execution. Variations at scale. Pattern application. The things designers used to spend 40 to 50 percent of their time on.</p><p>That work is now cheap. Which means competing on it makes you cheap.</p><p>When anyone can generate twenty concepts in ten minutes, concept generation is no longer premium. When any product manager can mock up a flow before the meeting, fidelity is no longer the proof of expertise. When the tool does the production, the person running the tool is overhead.</p><p>What becomes scarce — and scarcity is where leverage lives — is judgement.</p><p>What to pursue. What to kill. What not to build at all. When to move fast and when moving fast is the most expensive decision in the room. How to walk into a brief that says “use AI to improve trust” when nobody in the room can define trust, and move the work forward anyway.</p><p>That kind of judgement cannot be prompted. It cannot be templated. It does not get cheaper as the models improve.</p><p>And it lives in four specific skills that design has always claimed to value — and largely underdeveloped.</p><p><strong>Ambiguity tolerance.</strong> The brief is incomplete. The stakeholders disagree. The problem isn’t defined well enough to solve yet. You move forward anyway — not recklessly, but deliberately. You define the thing nobody else wanted to define, because leaving it undefined was more expensive than the conversation.</p><p><strong>Systems thinking.</strong> The thing you’re designing changes more than the UI. It changes the workflows, the staffing, the legal review, the data contracts, the things that break before they visibly break. You hold the whole system in your head while everyone else is focused on the feature.</p><p><strong>Risk interpretation.</strong> The feature technically works. You see how one wrong answer at scale becomes a headline, a support ticket avalanche, a trust problem that takes years to repair. You are the person who asks <em>what happens when this fails</em> while it is still being built.</p><p><strong>Organizational alignment.</strong> You get product, engineering, legal, and leadership pointed in the same direction before anything goes live, so nobody is surprised when it does. Not because you are a skilled meeting facilitator. Because you understood the political architecture of the organization and you navigated it on purpose.</p><p>These are not soft skills. They are not the nice-to-haves at the end of the curriculum after the Figma tutorials.</p><p>These are the skills that don’t get automated.</p><p>These are the skills that don’t get cheap.</p><p>And these are the skills that move you from interface altitude — to influence.</p><h3>AI just gave you 40% of your time back.</h3><p>You can look at what has been stripped away and say: <em>40% of my job got automated. What am I going to do.</em></p><p>Or you can say: <em>I just got 40% of my time back. What am I going to build with it.</em></p><p>Those are not the same question. And which one you ask will determine what the next decade of your career looks like.</p><p>Most designers, when automation accelerates, respond in one of three ways. They try to get faster. They try to master every new tool. They try to out-produce the machine. I understand the instinct. It is the only move the last decade trained us to make.</p><p>But you cannot win an efficiency war against automation. The harder you run on that track, the more replaceable you become. The machine will always be faster. That is not a temporary gap. That is the point.</p><p>The time you got back is not a buffer. It is not a grace period before the next disruption. It is the first real investment window most designers have had in a decade — unrushed, non-sprint, not already spoken for by Thursday’s handoff.</p><p>What you put into it is a choice. Not a passive one.</p><p>Design said for thirty years it wasn’t about how something looks — it was about how something works. That designers were builders, not decorators.</p><p>The scaffold we hid behind is no longer load-bearing.</p><p>The crossroads is not coming.</p><p>You’re already standing in it.</p><p><a href="http://www.jessaparette.com/"><em>Jessa Parette</em></a><em> is Head of Design at Byte by Yum! and an advisor to design and product leadership teams. She works at the intersection of AI-driven systems, organizational complexity, and experience design at scale. Speaking and advisory: </em><a href="http://www.jessaparette.com/"><em>www.jessaparette.com</em></a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=2b94bf2f5cae" width="1" height="1" alt=""><hr><p><a href="https://uxdesign.cc/ai-just-called-designs-bluff-2b94bf2f5cae">AI just called design’s bluff</a> was originally published in <a href="https://uxdesign.cc/">UX Collective</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[When the User Experience Happens Without the User]]></title>
            <link>https://medium.com/@jessa.parette/when-the-user-experience-happens-without-the-user-b117e2debd04?source=rss-1bd4a89f34e4------2</link>
            <guid isPermaLink="false">https://medium.com/p/b117e2debd04</guid>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[ux]]></category>
            <category><![CDATA[product-management]]></category>
            <category><![CDATA[design]]></category>
            <dc:creator><![CDATA[Jessa Parette]]></dc:creator>
            <pubDate>Tue, 14 Apr 2026 12:11:01 GMT</pubDate>
            <atom:updated>2026-04-14T12:11:01.356Z</atom:updated>
            <content:encoded><![CDATA[<h4>A framework for designing AI systems that earn trust without a visible interface</h4><h3>Nobody Is Designing the Part That Actually Matters</h3><p>Let me tell you what most AI product teams are doing right now.</p><p>They are obsessing over the interface.</p><p>The button copy. The loading state. The empty state illustration. The onboarding flow. Running usability tests on screens a user will touch for maybe thirty seconds — while the algorithm behind those screens makes a decision that affects that person’s career, their finances, their health, or their sense of whether they belong somewhere.</p><p>Nobody in the room is designing that part.</p><p>I’ve spent years working on AI-driven systems at enterprise scale. Products used by tens of millions of people daily. Algorithmic platforms matching hundreds of thousands of employees. Systems where a single design decision made upstream — in the data layer, in the signal definition, in the success metric nobody questioned — quietly shapes what’s possible for every user downstream.</p><p>I have been in those rooms.</p><p>I have been the person asking the question nobody wanted to answer.</p><p>Most teams treat invisible experience as an engineering problem. Build a better model. Clean the data. Improve the score. Ship it.</p><p>It isn’t an engineering problem.</p><blockquote>It’s a trust problem. And trust has to be designed before the system is built — not after it ships and someone finally notices something is wrong.</blockquote><p>Here is the framework I wish I’d had going in. Five questions. Each one ends with a design decision — not a reflection exercise, not a workshop sticky note. An actual thing you go do.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*xkSEgL5ge-Xj8Ynw" /><figcaption>Photo by <a href="https://unsplash.com/@markusspiske?utm_source=medium&amp;utm_medium=referral">Markus Spiske</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><h3>Question 1 — “Are we using the right data, or just the available data?”</h3><h4>→ Redefine the inputs. Before the model trains.</h4><p>Most AI systems are built on whatever data is already lying around.</p><p>Nobody sat in a room and decided that job title, tenure, and location were the right signals. They were just there. So they got used. The model trained on them. The product shipped. And somewhere in the gap between <em>available</em> and <em>right</em>, a few thousand people got results that had nothing to do with what they actually needed.</p><p>That gap is a design problem.</p><p>It just doesn’t look like one yet.</p><p>Job title. Tenure. Location. These signals are available. They are also, frequently, terrible — proxies for what you actually care about that flatten nuance, encode decades of organizational history, and produce results that are technically defensible and experientially wrong.</p><p>I learned this building a mentorship matching platform inside a massive organization. We had more data than mosquitoes at a blood bank. Petabytes of it. Rich, layered, deeply human records about hundreds of thousands of people.</p><p>The early model kept surfacing the same kinds of mentors. Senior. Long-tenured. Well-documented. The people who had been around long enough to leave a trail in the system.</p><p>The algorithm was working exactly as designed.</p><p>It was also completely missing the point.</p><p>Research we’d done with actual mentors and mentees told us what actually predicted a good match: learning style, communication preferences, what someone <em>wanted</em> from the relationship rather than what they’d already accomplished. The available signals were answering a completely different question than the one we were asking.</p><p>So we went back to the data team and redefined what inputs the model was allowed to use.</p><p>That conversation was not comfortable. The model had already been partially trained. There were timeline implications. People had to redo work.</p><p>We did it anyway.</p><blockquote><strong>The question to bring to your next data meeting:</strong> If we stripped out every proxy signal and kept only what directly predicts the outcome we actually care about — what would the model have left?</blockquote><p>If the answer is “not much” — that’s not a reason to keep the proxies. That’s a reason to do the research first.</p><h3>Question 2 — “Who is already invisible in this data — and does our system know that?”</h3><h4>→ Enforce the correction upstream. Hide it from the UI.</h4><p>This is the part of the AI conversation that makes rooms uncomfortable.</p><p>Not because it’s complicated. Because it requires naming something organizations would rather not name.</p><p>The data your algorithm trains on reflects the history of whoever built the systems that collected it. Most large American enterprises built between the 1960s and 1990s hired predominantly men. That history didn’t disappear when you decided to build something new on top of old infrastructure. It calcified — into job descriptions written in male-dominant language, into performance qualifiers tracking traditionally male career patterns, into data structures reflecting who was historically considered worth measuring and promoting and documenting.</p><p>Your algorithm inherits all of it.</p><p>And then it reproduces it. With mathematical confidence. And no visible fingerprints.</p><p>Before we even get to that: data quality is its own disaster. Ingesting and reconciling dirty, incomplete entries eliminated more than half our potential user base before we matched a single person. Not a rounding error. More than half. Gone before we started, because the data that was supposed to represent these people was riddled with duplicates, missing fields, and inconsistencies across business units that had never needed to talk to each other.</p><p>Data abundance is not data quality.</p><p>Confidence without accuracy isn’t a feature. It’s how you fail people at scale without ever finding out.</p><p>We fixed the surface layer — removing names, stripping photos, deprioritizing location. UI decisions. They helped at the margins.</p><p>Then I pushed further.</p><p>I worked with the data team to enforce gender-equal candidate surfacing directly in the algorithm. Not as a UI filter. Not as a toggle. Not as anything a user could see or adjust. A hidden correction built into the system logic, actively working against the imbalance the organization’s own history had created.</p><p>Some people called this artificial.</p><p>I called it corrective. The data was already artificial — artificially skewed by decades of decisions nobody was making anymore but that the system was still faithfully reproducing.</p><p>You cannot just be unbiased.</p><p>Unbiased is passive. A shrug. “We didn’t mean to.” Anti-bias is a verb. It means going looking for who the system is quietly failing before anyone reports it. Asking out loud, in a room where it’s inconvenient: <em>is that fair?</em> And not accepting “the algorithm did it” as an answer.</p><blockquote><strong>The question to ask before your next model review:</strong> If we ran our results broken out by gender, race, age, and tenure — would they be equal?</blockquote><p>And if they’re not: is that the world as it should be, or the world as it has been?</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*PtMqW5vKVCQOTWBX" /><figcaption>Photo by <a href="https://unsplash.com/@jontyson?utm_source=medium&amp;utm_medium=referral">Jon Tyson</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><h3>Question 3 — “Is our success metric good for the user, or just convenient for us?”</h3><h4>→ Rewrite the metric. Then defend it when everyone pushes back.</h4><p>Every AI system has a success metric.</p><p>That metric is a design decision. Most organizations don’t treat it that way.</p><p>They treat it like a math problem — something the data science team owns, expressed in a language designers don’t speak, decided in a meeting designers weren’t invited to. The product gets built around that number. Nobody asks whether optimizing for it actually makes someone’s life better.</p><p>An AI that maximizes match volume is not the same thing as an AI that creates meaningful connections.</p><p>Those are different design decisions disguised as the same dashboard.</p><p>Match volume is easy to instrument. Meaningful connection is messy, slow, and harder to put in a quarterly review. So most teams pick match volume, ship the product, declare success — while the actual experience quietly fails the people it was supposed to serve.</p><p>Nobody notices because the number looks great.</p><blockquote><strong>The question to ask in the next roadmap meeting:</strong> If our system achieved its success metric perfectly — would a real person’s life actually be better?</blockquote><p>Or would we just have a great dashboard?</p><p>If you can’t answer that with a specific yes and a specific person in mind — the metric is wrong.</p><p>Rewrite it. This will create friction. The model is already optimized for the old metric. Engineering has instrumented it. Product has committed to it. You are now the person asking everyone to change something that was already working.</p><p>Do it anyway.</p><p>The metric is the design. Everything downstream is just execution.</p><h3>Question 4 — “Does the user feel understood — or just processed?”</h3><h4>→ Design the signal, not the explanation.</h4><p>Here is the mistake I see most often in AI product work.</p><p>Teams trying to build trust by explaining the algorithm.</p><p><em>“Here’s why we recommended this.” “Our model uses 47 signals.” “Click here to learn how our AI works.”</em></p><p>Nobody reads that.</p><p>And more importantly — that’s not what trust is built on.</p><p>Think about the last time you trusted someone. You didn’t trust them because they explained their decision-making process to you. You trusted them because they said something that made you feel like they actually understood your situation. Like they were paying attention to <em>you</em> specifically. Not running you through a process.</p><p>AI systems can do this.</p><p>Most don’t. Because the team building them is focused on accuracy and explainability instead of the simpler question: <em>does this feel personal?</em></p><p>You don’t need to explain the algorithm to build trust. You need to demonstrate that the system understands the person.</p><p>Those are completely different problems. One is a transparency problem — hard, legally complicated, often impossible at the model level. The other is a design problem — solvable, if you’ve done the first three questions correctly.</p><p>If you’ve defined the right signals, corrected for bias, and built toward an outcome that actually serves the user — the trust work becomes about surfacing <em>why this result, for this person, right now</em> in a way that feels specific rather than generated.</p><p>Not “our algorithm matched you.”</p><p>But “we think you’d connect well here because of where you want to go — not just where you’ve been.”</p><blockquote><strong>The question to sit with:</strong> If the user could see exactly what this system knows about them — would they feel understood?</blockquote><p>Or would they feel like a data point?</p><p>Design toward the former. Starting with the inputs you defined in Question 1.</p><h3>Question 5 — “When this fails someone, will we even know?”</h3><h4>→ Build the failure signal before you need it. You will need it.</h4><p>In traditional UX, failure is visible.</p><p>A broken flow. A dead end. A confusing error state. Users see it, they tell you, you fix it. The feedback loop exists because the interface exists.</p><p>In invisible systems, failure is silent.</p><p>The wrong match. The recommendation that lands flat. The routing decision that felt subtly off — in a way the user couldn’t name — so they just stopped. Quietly. Without a ticket, without a complaint, without anything that shows up in the data.</p><p>The dashboard stays green. The metric holds. Nobody asks questions.</p><p>And somewhere out there, a person stopped trusting the system.</p><p>The system never found out.</p><p>Silent failure is the worst kind because nobody sounds the alarm. And there’s a question that almost never gets asked in the design review: <em>when this fails someone in a consequential way — who owns that?</em></p><p>Not the model. Not the data. A person. A person who was in the room when the design decisions were made. In most organizations, that person should be a designer.</p><p>Which means you have to build the mechanism for finding out before you ship.</p><p>Not after a user complains. Not after a pattern shows up in the churn data six months later.</p><p>Before.</p><p>Make the failure signal frictionless — a single tap, a quick rating, a low-cost way to say <em>“this wasn’t right.”</em> Close the loop so the signal feeds back into the system. Make someone accountable for reading it.</p><blockquote><strong>The question to answer before you launch:</strong> Right now, if this system failed someone in a consequential way — how would we find out?</blockquote><p>And what could they actually do about it?</p><p>If the answer involves emailing a support address — you don’t have an answer.</p><h3>When It’s All Said and Done</h3><p>The future of experience design is not more intuitive interfaces.</p><p>It’s designing experiences where the interface disappears — and the design work moves upstream. Into the system logic. Into the data decisions. Into the success metric nobody questioned because it looked like a math problem.</p><p>These are design problems.</p><p>They don’t look like design problems yet because designers haven’t insisted on being in the room where they get made.</p><p>That is the work.</p><p>Get in the room before the model trains. Ask the question everyone is hoping nobody asks. Be the annoying designer — I promise it’s better than the alternative, which is shipping something that fails people quietly, confidently, and at scale.</p><h4>The interface is the easy part.</h4><h4>Do the hard part first.</h4><p><em>I write about designing AI systems that earn trust, leading through organizational complexity, and the things design leadership books skip. Follow me here or find me at jessaparette.com.</em></p><p><em>I work with design and product leadership teams on exactly these problems — speaking, workshops, and advisory. If your organization is building AI-driven experiences and needs someone to ask the inconvenient questions, reach out.</em></p><p><em>Jessa Parette is Head of Design at Byte by Yum! and an advisor to design and product leadership teams. </em><a href="https://www.jessaparette.com/"><em>jessaparette.com</em></a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=b117e2debd04" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How We Built Tinder for Mentors — Inside a Fortune 1 Retailer]]></title>
            <link>https://medium.com/@jessa.parette/how-we-built-tinder-for-mentors-inside-a-fortune-1-retailer-e857fbb2cfa7?source=rss-1bd4a89f34e4------2</link>
            <guid isPermaLink="false">https://medium.com/p/e857fbb2cfa7</guid>
            <category><![CDATA[diversity-and-inclusion]]></category>
            <category><![CDATA[design]]></category>
            <category><![CDATA[ux]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[product-management]]></category>
            <dc:creator><![CDATA[Jessa Parette]]></dc:creator>
            <pubDate>Mon, 06 Apr 2026 12:31:01 GMT</pubDate>
            <atom:updated>2026-04-07T14:52:02.707Z</atom:updated>
            <content:encoded><![CDATA[<h4>We had petabytes of data, a swipe-right matching system, and results that kept coming back mostly men. Here’s what happened when I wouldn’t let that go.</h4><h4>“What If It Was Like Tinder?”</h4><p>The room got quiet in the way rooms get quiet when someone says something that’s either brilliant or career-ending and nobody’s sure which yet.</p><p>I was a brand new design leader — still carrying an IC design workload, still figuring out what “leading” actually meant in a company this size — when the data science team walked in with an unusual problem. They had petabytes of employee data. Work history. Tenure. Location. Performance signals. More data than most organizations will ever touch. And they had no idea what to build with it.</p><p>They looked at me and asked the best and worst question a designer can receive:</p><p><em>“What should this be?”</em></p><p>I did not have a visionary answer ready. I did not have a framework on a whiteboard or a three-year roadmap in my back pocket. What I had was a team of two, a conference room, and a half-formed thought that kept coming back to me: <em>the hardest part of finding a mentor at a massive company isn’t the mentorship. It’s the finding.</em></p><p>So I did what any reasonable designer does when they don’t know the answer yet.</p><h3>We Started With Tinder</h3><p>Before running a single focus group or writing a single user story, I sat down with my team of two and did something that raised a few eyebrows: we ran a competitive SWOT analysis on two products that had nothing to do with enterprise HR software.</p><p><strong>Tinder.</strong> And <strong>Good &amp; Co.</strong></p><p>Tinder is a masterclass in impulse matching — low friction, fast signal, swipe-based decisioning that removes the cognitive overhead from connection. Good &amp; Co. was a workplace culture platform built around compatibility between people and organizations.</p><p>Neither was a mentorship product. Both were secretly solving the same problem we’d been handed: <em>how do you connect the right people with the minimum amount of friction?</em></p><p>The SWOT forced us to ask: what works about each of these? What fails? What would an enterprise version of their best qualities look like — and what would we need to leave behind?</p><p>That exercise gave us a hypothesis before we’d talked to a single user: <strong>the connection mechanism is the product. Not the mentorship itself. Just the match.</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*kBZum_LgzuVExRHu" /><figcaption>Photo by <a href="https://unsplash.com/@ingoodhands?utm_source=medium&amp;utm_medium=referral">Good Faces Agency</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><h3>Then We Talked to People</h3><p>Hypothesis in hand, we ran focus groups — separately, with mentors and mentees across the organization. The questions were simple on purpose:</p><ul><li>What do you look for in a mentor?</li><li>Is it hard to find one today?</li><li>Why?</li></ul><p>We paired qualitative questions with Likert scale items. Something to measure, not just something to feel.</p><p>What came back was consistent. Clarifying. And a little humbling.</p><p>The problem wasn’t mentorship. People understood mentorship. They wanted it. They knew exactly what a good one looked like. The problem was finding someone outside their immediate corner of the building.</p><p>In a company of hundreds of thousands of people spread across every geography, function, and level — most employees were invisible to each other. Mentor programs had been attempted before. They kept failing for the same reason: <em>you can only find who you can already see.</em></p><p><em>We didn’t need to make mentorship better. We needed to make discovery easier.</em></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*cJoIXs-8C7tW7wvu" /><figcaption>Photo by <a href="https://unsplash.com/@bekkybekks?utm_source=medium&amp;utm_medium=referral">Bekky Bekks</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><h3>We Had Everything We Needed. Except Clean Data.</h3><p>Here’s where most product stories skip the hard part.</p><p>We had petabytes of employee data. On paper, everything a matching algorithm could want. In practice, ingesting and reconciling dirty, incomplete, and inconsistent entries eliminated more than half of our potential user base before we matched a single person.</p><p>Duplicate records. Missing fields. Job titles formatted seventeen different ways across different business units. Data entered by humans, maintained inconsistently, and never audited for the purpose of powering a product.</p><p>Data abundance is not data quality. More data doesn’t mean better outcomes — it means more <em>confident</em> outcomes. Which is far more dangerous when the underlying data is wrong.</p><p>We learned that the hard way.</p><h3>“Wait. Is That Fair?”</h3><p>We were deep into testing when I noticed something.</p><p>The results kept coming back mostly men. Senior men. Decorated men. Men whose career trajectories were long and well-documented in the system. By every available signal, the algorithm was working exactly as designed.</p><p>I looked at the results and asked a question that made the room uncomfortable:</p><p><em>“Is that fair?”</em></p><p>Not “is it accurate?” Not “is it statistically defensible?” Not “does it meet the success metric?”</p><p><em>Is it fair?</em></p><p>The data science team had an answer for the first three questions. Nobody had a clean answer for the fourth.</p><p>Here’s what the data was actually telling us: this organization, like most large American companies built between the 1960s and 1990s, had hired predominantly men. That history hadn’t disappeared when we started building. It had calcified — into job descriptions written in male-dominant language, into performance qualifiers that tracked traditionally male career patterns, into data structures that reflected who had historically been considered worth measuring and promoting. The algorithm wasn’t broken. It was telling the truth about the company’s history.</p><p>And in doing so, it was about to reproduce it.</p><p><strong>I wouldn’t let it go</strong>. I kept asking the question in rooms where it was inconvenient to ask it. I pushed it into design reviews. I pushed it into data conversations. At one point, I pushed it hard enough that legal got involved — not because anyone had done anything wrong, but because the annoying designer in the room kept insisting that “the system works” was not the same as “the system is fair,” and eventually that distinction needed an actual answer.</p><p>We addressed the surface layer first. First name and last initial only — no full names that signal gender or ethnicity. Randomized abstract icons instead of photos — no appearance, no presentation, no visual gender cues. Location deprioritized as a primary match signal.</p><p>These were UI decisions.</p><p>The bias wasn’t living in the UI.</p><p>So I pushed further. I worked with the data team to enforce gender-equal candidate surfacing directly in the algorithm — not as a filter, not as a toggle, not as anything a user could see or adjust. A hidden correction built into the system logic, actively working against the imbalance the organization’s own history had created.</p><p>There’s a word for what I was doing in those rooms, asking the same question nobody wanted to answer.</p><p><em>Annoying.</em></p><p>There’s a better word too.</p><p><strong><em>Anti-bias.</em></strong></p><p>Not unbiased — that’s passive, a shrug, a “we didn’t mean to.” Anti-bias is a verb. It means you went looking for the problem before it found you. It means you asked <em>“is that fair?”</em> in the meeting where everyone else had already moved on. It means you made legal nervous and the data team frustrated and the timeline longer.</p><p>It means you didn’t stop until the answer was actually yes.</p><h3>What We Built</h3><p>The product that came out of this process was simple by design. Employees could be matched with potential mentors anywhere in the organization — across geography, hierarchy, and function — based on work history, learning goals, and communication preferences rather than proximity and org chart position.</p><p>Swipe right. Swipe left. Tinder for Mentors.</p><p>Match volume increased. Prediction accuracy improved over time as the feedback loop fed back into the algorithm. People were connecting with mentors they never would have found through any formal program — across the country, across functions, across every level of the organization.</p><p>The product was later highlighted by the company’s Chief Inclusion Officer and publicly announced at SXSW as a first-of-its-kind mentorship system — and a foundational example of what human-centered algorithmic design can look like at enterprise scale.</p><h3>What This Story Is Actually About</h3><p>It’s not about the company. It’s not really even about the algorithm.</p><p>It’s about what happens when design is in the room at the beginning — when a data science team asks “what should this be?” and a designer actually answers with research, not assumptions.</p><p>It’s about the value of running a SWOT on Tinder when you’re building enterprise HR software. It’s about focus groups that replace gut instinct with evidence. It’s about the unglamorous work of data quality that determines whether a product reaches half its intended users or all of them.</p><p>And it’s about the design decisions that never appear in the interface — the hidden corrections, the enforced equalities, the values encoded in system logic — that determine whether a product reproduces the world as it has been, or quietly helps build the world as it should be.</p><p>Those decisions don’t make the changelog. They don’t show up in the UI. Most design reviews never touch them.</p><p>But they are the most consequential design work happening in AI right now.</p><p>And almost nobody is asking <em>“is that fair?”</em></p><h3>The Book That Said It First — About Something Else</h3><p>In <a href="https://www.ibramxkendi.com/how-to-be-an-antiracist"><em>How to Be an Antiracist</em>, Ibram X. Kendi </a>makes an argument that reframed an entire cultural conversation: it is not enough to simply not be racist. There is no neutral position. The systems already have racism built into them — in policy, in hiring, in data, in who gets resourced and who gets overlooked. Passivity doesn’t opt you out of that system. It just means you stop noticing it while it keeps running.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*JGVgOBMbWO4oGG3i" /><figcaption>Photo by <a href="https://unsplash.com/@timwildsmith?utm_source=medium&amp;utm_medium=referral">Tim Wildsmith</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><p>His argument: if you are not actively antiracist, you are — by default — participating in the reproduction of racist outcomes, regardless of your intentions.</p><p>I am not here to make a statement about race. Kendi already made it, better than I could.</p><p>What I am here to say is that the same logic applies — with identical precision — to bias in AI systems.</p><p>It is not enough to take the unconscious bias training. It is not enough to remove the protected attributes from your model inputs. It is not enough to say “we didn’t intend to discriminate.” The systems already have bias built into them — in the data they were trained on, in the signals they were designed to optimize, in the decades of hiring decisions and organizational choices that shaped the information your algorithm now treats as ground truth.</p><p>Passivity doesn’t opt your AI system out of that history. It just means you stop looking for where the bias lives while the system keeps running.</p><p><strong>To build AI systems that are genuinely fair, you cannot simply be unbiased. You have to be actively anti-bias.</strong></p><p>That means going looking for the problem before anyone reports it. It means asking not just “did we discriminate?” but “who is this system quietly failing while the metrics look fine?” It means making deliberate, sometimes hidden interventions — not as a feature, but as a correction for what the data got wrong.</p><p>It means being the annoying designer in the room who keeps asking <em>“but is that fair?”</em> long after everyone else has moved on.</p><p>It is not comfortable work. It won’t show up in your sprint velocity. It will occasionally make legal nervous. It may get you fired.</p><p>It is also the only way to build AI systems that deserve the trust of the people they affect.</p><p><em>If this resonated, I write about designing AI systems that earn trust, leading through organizational complexity, and the things design leadership books don’t tell you. Follow me here on Medium or find me at jessaparette.com.</em></p><p><em>I work with design and product leadership teams on exactly these problems — speaking, workshops, and advisory. If your organization is building AI-driven experiences and wants to think more carefully about what you’re actually building, reach out.</em></p><p><em>Jessa Parette is Head of Design at Byte by Yum! and an advisor to design and product leadership teams. She works at the intersection of AI-driven systems, organizational complexity, and experience design at scale. Speaking and advisory: jessaparette.com</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=e857fbb2cfa7" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Why Voice AI Still Feels Wrong (Part 1): Discovering the Invisible Layer of Conversation]]></title>
            <link>https://medium.com/@jessa.parette/designing-for-porosity-the-missing-layer-in-voice-ai-c53016a93dc5?source=rss-1bd4a89f34e4------2</link>
            <guid isPermaLink="false">https://medium.com/p/c53016a93dc5</guid>
            <category><![CDATA[product-management]]></category>
            <category><![CDATA[design]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[voice-technology]]></category>
            <category><![CDATA[ux]]></category>
            <dc:creator><![CDATA[Jessa Parette]]></dc:creator>
            <pubDate>Wed, 01 Apr 2026 10:26:01 GMT</pubDate>
            <atom:updated>2026-07-17T08:01:22.696Z</atom:updated>
            <content:encoded><![CDATA[<h4>Voice AI keeps getting smarter. It keeps sounding wrong. The problem isn’t the script — it’s the invisible layer we haven’t designed for yet.</h4><p>There is a particular kind of frustration that comes from sitting in a room full of smart people trying to design voice AI by feel.</p><p>The engineers are capable. The dialog flows technically work. And still, something is off — the system sounds cold, or slightly robotic, or wrong in a way no one in the room can precisely name. So the conversation circles. People adjust the script. Someone suggests a different word choice. The problem doesn’t move.</p><p>I came to this problem through my work as Head of Design at Byte by Yum! — the AI platform behind some of the largest restaurant brands in the world. Voice AI isn’t hypothetical in my world. It’s on the table right now, in real products, with real stakes.</p><p>And I think it’s about to be on everyone’s table.</p><p>Design leaders are already deep in the AI conversation — copilots, generative interfaces, AI-assisted workflows. That’s where most of the energy is. But voice AI is the one that will sneak up on us. Because it doesn’t look like a design problem until it already is one. There’s no screen to critique, no layout to redline. By the time it feels wrong to a user, the system is already in production — and no one can quite explain why.</p><p>The issue is never the words. It never is.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*eGIH6_2HPlvBqNBwwKI2KA.jpeg" /></figure><h3>The Problem Isn’t the Script</h3><p>For decades, design has codified standards for visible interfaces. Grids, color systems, typography, accessibility rules. These frameworks exist because practitioners recognized that shared principles — not individual intuition — were what separated craft from guesswork.</p><p>Voice AI has no such foundation.</p><p>Most implementations are built around dialog flows and intent classification. That is necessary. It is not sufficient. To understand why, try a simple demonstration. One sentence. Five words.</p><p><em>“I didn’t take her wallet.”</em></p><p>Say it out loud. Now say it again, stressing a different word each time.</p><p>Stress <em>I</em> — and you’re denying personal involvement. Someone else did it. Stress <em>didn’t</em> — a flat denial of the act itself. Stress <em>take</em> — maybe you borrowed it. Stress <em>her</em> — it was the ownership that mattered, not the act. Stress <em>wallet</em> — you took something, just not that.</p><p>Same five words. Five completely different meanings.</p><p>This is the hidden physics of conversation — the way tone, lilt, and timing change intent before a single syntactic element shifts. And this is exactly where most voice AI fails. It hears the words. It misses the meaning.</p><p>Most teams respond to this problem by rewriting the script. More words, better words, cleaner words. But the script was never the problem. The problem is that we’ve been designing for language when we should be designing for something that exists beneath it.</p><h3>What Porosity Is</h3><p>The word I’ve landed on for that something is <strong>porosity</strong>.</p><p>Porosity is the patterns of rhythm and sound that carry meaning between and beneath the words. It is the invisible layer of conversation — the part that operates before language, beyond language, and sometimes instead of language entirely.</p><p>Humans are wired to parse meaning from tone and timing automatically, faster than conscious thought, across every language and culture. We hold entire conversations without words — just sounds. A sigh signals frustration. A gasp signals surprise. A long pause signals something is wrong. None of that requires a sentence. The meaning arrives before the language does.</p><p>Think about the last time you answered the phone and knew, from a single “hello,” that something was off. No content. No context. Just a sound — and you knew.</p><p>That is porosity. And most voice AI systems are completely deaf to it.</p><p>The question for designers is not how to write better scripts. It is how to teach a system to operate in this invisible layer of sound and silence. That is a fundamentally different design problem. It requires different inputs, different metrics, different prototyping methods, and a different vocabulary for talking about what we’re building.</p><h3>What We’re Actually Designing For</h3><p>When I gave this talk at the Design Leadership Summit, I walked through what designing for porosity actually means in practice. It is not one thing. It is four.</p><p><strong>Lilt.</strong> The musical quality of a phrase — the rise and fall that makes a sentence feel like a question, a reassurance, or a command even before the words land. Lilt is what makes a customer service voice feel warm versus procedural. It is entirely absent from most voice AI outputs, which tend toward a flat, declarative rhythm that feels efficient and nothing else.</p><p><strong>Silence.</strong> The gaps between turns that human brains read as social signals. A pause of the right length signals thinking. A pause that runs too long signals hesitation, confusion, or rejection. Silence is social glue — until it becomes social distance. In voice AI, silence is almost always treated as a technical problem to minimize rather than a design variable to calibrate.</p><p><strong>Stress.</strong> The weight placed on individual words within a phrase. As the wallet sentence shows, stress is not decoration — it is meaning. A system that produces correct words in a flat, unstressed delivery is not communicating. It is reciting.</p><p><strong>Tone.</strong> The acoustic quality that tells someone, within seconds, whether they are trusted, rushed, heard, or dismissed. Tone operates below the level of content. It is what users are responding to when they describe a system as “cold” or “robotic” — and it is almost never explicitly designed.</p><p>Without porosity, a voice AI can’t hear what we really mean. It can only hear what we literally said. And without designing for porosity, a voice AI can’t <em>sound</em> like it understands — regardless of how accurate its responses are.</p><h3>Why This Is a Design Leadership Problem</h3><p>It would be convenient if porosity were purely a technical problem — something for NLP engineers to solve with better models. It is not.</p><p>The decisions that determine whether a voice system feels human are design decisions: what emotional signals to detect, how silence should be calibrated, what tone the system projects in different contexts, how recovery should sound when something goes wrong. These require design judgment, design standards, and design governance. They require someone in the room asking not just “does it work?” but “does it <em>feel</em> right — and do we know why?”</p><p>Right now, most organizations don’t have that person in that room.</p><p>Voice AI is moving fast. The teams building these systems are capable and well-intentioned. But without a design framework for the invisible layer, they are optimizing for accuracy while the user experience degrades at the level of relationship. Users don’t file bug reports about porosity. They just stop trusting the system. They call it robotic. They disengage. And by the time the data surfaces, the damage is done.</p><p>Design leaders need to get ahead of this. Not because voice AI is new — it isn’t — but because the standards for designing it don’t exist yet. We have decades of practice designing for screens. We are starting from scratch on sound.</p><h3>What Comes Next</h3><p>Understanding porosity changes how you see voice AI failures. The script isn’t the problem. The flows aren’t the problem. What’s missing is a set of governing principles for the invisible layer — for the physics of spoken interaction that operates whether your system accounts for it or not.</p><p>In my next article, I’ll introduce the framework I developed to address that gap: <strong>The Five Laws of Invisible Conversation.</strong> These are not tips or tactics. They are the structural conditions under which human conversation functions — grounded in linguistic, psychological, and conversational science. They are, as best I can define them, the laws that govern the invisible layer.</p><p>Once you understand porosity, you can start to name what it requires.</p><p>That’s where we’re going.</p><p><em>Jessa Parette is Head of Design at Byte by Yum! and a speaker on Voice AI, design systems, and the future of invisible interfaces. This piece is adapted from a talk originally presented at the Design Leadership Summit, 2025.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=c53016a93dc5" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>