<?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 Apto Solutions on Medium]]></title>
        <description><![CDATA[Stories by Apto Solutions on Medium]]></description>
        <link>https://medium.com/@apto.solutions?source=rss-1adf5dd40347------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*MQbTJijS6wgCNq5qkgH9mw.png</url>
            <title>Stories by Apto Solutions on Medium</title>
            <link>https://medium.com/@apto.solutions?source=rss-1adf5dd40347------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Sat, 21 Nov 2026 01:19:22 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@apto.solutions/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[AI agents and your telemetry: what it costs and what your platform needs]]></title>
            <link>https://medium.com/@apto.solutions/ai-agents-and-your-telemetry-what-it-costs-and-what-your-platform-needs-c477f4f616da?source=rss-1adf5dd40347------2</link>
            <guid isPermaLink="false">https://medium.com/p/c477f4f616da</guid>
            <category><![CDATA[cribl]]></category>
            <category><![CDATA[cost]]></category>
            <category><![CDATA[telemetry]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[criblcon]]></category>
            <dc:creator><![CDATA[Apto Solutions]]></dc:creator>
            <pubDate>Thu, 19 Nov 2026 10:31:01 GMT</pubDate>
            <atom:updated>2026-11-19T10:31:01.578Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*CvcwUfqXLqsP0dNzXXxoyA.png" /></figure><p><em>The three meters every telemetry budget now has, and six tests for an agent-ready platform.</em></p><h3>1. The three meters: storage, query and now inference</h3><p><a href="https://www.youtube.com/watch?v=WnVfUH7xGm4"><strong><em>WATCH IN THE CRIBLCON 2026 KEYNOTE</em></strong></a></p><p><strong>7:08 Machine-speed queries, 10x load</strong></p><p><strong>1:14:30 Three meters explained</strong></p><p><strong>1:15:46 Free inference announcement</strong></p><p>Telemetry budgets used to have two meters. From this year they have three, and the third is the one your finance director has not seen yet.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*l3EUedtBtITTTKva.png" /></figure><p><em>Figure 1 — This figure shows the three meters a telemetry budget now has; inference is new and has no spending history.</em></p><h4>1.1 The two you know</h4><p>For a decade, the cost of a logging or security platform has come down to two questions: how much data do we keep, and how much compute do we need to search it. Cribl’s long-running argument is that data grows at about 30% a year and budgets do not. Cribl’s CEO, Clint Sharp told the CriblCon audience he has asked thousands of customers whether their budget is growing at 30% and has never had a yes.</p><h4>1.2 What agents do to the second meter</h4><p>AI agents are now being pointed at this data to investigate incidents. A person runs a query, reads the result and thinks. An agent tests many hypotheses in parallel and never stops for coffee. Sharp’s estimate is that agents issue around <a href="https://www.youtube.com/watch?v=WnVfUH7xGm4&amp;t=428s">ten times the queries humans do</a>. A platform sized for analysts will not hold up, and one priced per search will become very expensive very quickly.</p><p>There is a second effect. Data that was not worth keeping, because no human had time to look at it, becomes worth keeping once an agent can. So the first meter rises too.</p><h4>1.3 The third meter</h4><p>Every question an agent asks consumes tokens, the unit AI providers bill in. Sharp called this <a href="https://www.youtube.com/watch?v=WnVfUH7xGm4&amp;t=4470s">“the next big fight with your CFO”</a>: a spend category with no history, no forecast and no obvious owner.</p><p>Cribl published research alongside the event that makes the point sharply. It tested AI models on 30 real IT and security investigations and reports that accuracy varied by 17% across models while cost varied by 20 times. In other words, the most expensive model is rarely the right default. That benchmark is Cribl’s own and has not yet been independently reproduced, but the direction matches what most teams find when they look at their own usage.</p><h4>1.4 Cribl’s answer</h4><p>Two announcements address the third meter. StreamAI, due in Cribl.Cloud in November, is a gateway that sends each request to the cheapest model that can do the job, enforces budgets and records every call. And Sharp announced that <a href="https://www.youtube.com/watch?v=WnVfUH7xGm4&amp;t=4546s">inference for work done on Cribl’s platform is included</a>: “not a credit, not a trial”, and not a line to negotiate at renewal.</p><p>Read that carefully before building a business case on it. Cribl’s written announcement ties free inference to its router choosing the model. If you insist on a particular model, you pay that provider’s rate. No fair-use limits have been published. Cribl still charges for the infrastructure underneath, so heavier AI use is not free in total.</p><h4>1.5 Five questions for any telemetry or AI vendor</h4><ol><li>Which of the three meters do you charge on, and which are bundled?</li><li>If our query volume rises tenfold because of agents, what happens to our bill?</li><li>Who chooses the AI model, and what does it cost when we choose?</li><li>Can we see token spend by team and by use case?</li><li>What stops a runaway agent from spending a month’s budget overnight?</li></ol><p>The organisations that will be comfortable in two years are the ones that start measuring the third meter now, while it is still small.</p><p>Since the keynote: Cribl’s <a href="https://cribl.io/blog/introducing-secit-bench-cribls-ai-telemetry-benchmark-a-new-standard-for-evaluating-ai-models-in-it-and-security-workflows/">SecIT Bench write-up</a> gives the detail behind the headline. The top model averaged $3.27 per investigation, spend ranged from $0.28 to $7.34 per session, and the costliest set-ups were not the most accurate. Two cautions: the write-up covers 14 models where the launch press release says 20, and most scenarios are simulated.</p><h3>2. Agents are the new consumer of your telemetry: six things your platform needs</h3><p><a href="https://www.youtube.com/watch?v=WnVfUH7xGm4"><strong><em>WATCH IN THE CRIBLCON 2026 KEYNOTE</em></strong></a></p><p><strong>10:38 Six capabilities</strong></p><p><strong>12:25 Context: Jira, wiki, code</strong></p><p><strong>12:50 Governing agents and audit trail</strong></p><p>Your logs, metrics and traces were collected for people to read. Increasingly they will be read by software agents instead, and agents ask for different things. This article offers a six-point checklist for judging whether your current platform is ready, whoever supplies it.</p><p>The list comes from Cribl’s keynote at CriblCon 2026, where Clint Sharp set out <a href="https://www.youtube.com/watch?v=WnVfUH7xGm4&amp;t=638s">six capabilities</a> he argues an AI-era telemetry platform must have. Cribl naturally believes it scores six out of six. The questions are worth asking of any vendor, including Cribl.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*cQEFA5SY3tjSY60d.png" /></figure><p><em>Figure 2 — This figure shows the six tests for an agent-ready platform; governance (test 5) deserves the closest look.</em></p><h4>2.1 Why agents change the requirement</h4><p>An analyst under pressure needs a fast answer to one question. An agent explores dozens of possible causes at once, around the clock. Cribl’s chief technology officer described this as problem-solving at machine speed and put the extra load at roughly ten times what humans generate. A platform has to serve both: quick responses for people, and heavy parallel load for machines.</p><h4>2.2 The six tests</h4><ol><li>Open. Can you bring your own agent or AI model, and reach your data through open interfaces without being charged per call? If getting your own data out is metered, every agent you add raises the bill.</li><li>Full-fidelity retention. Can you afford to keep everything, for long enough? Investigations often look back nine to eighteen months. Data that was too dull for a human to review may be exactly what an agent needs.</li><li>Search where the data lives. Can the platform query data it does not hold, such as an archive in your own cloud storage or another vendor’s store? Some data cannot move, for cost or regulatory reasons, and an agent blind to it will reach wrong conclusions confidently.</li><li>Context. Telemetry says what happened. <a href="https://www.youtube.com/watch?v=WnVfUH7xGm4&amp;t=745s">Tickets, wikis and source code</a>say what was supposed to happen and who owns it. Can an agent reach that context alongside the logs?</li><li>Governance. This is the test most estates fail today. Can you <a href="https://www.youtube.com/watch?v=WnVfUH7xGm4&amp;t=770s">control exactly what each agent can reach</a>, and is there a full audit trail of what it did? Cribl reported at the event that around 1.5% of the events its scanner inspects contain sensitive data. An ungoverned agent with broad access to logs is a data-protection incident waiting to be written up.</li><li>Built-in inference. Is the AI compute part of the platform, or a separate contract with a separate bill? Neither answer is wrong, but you should know which you have and what it costs.</li></ol><h4>2.3 Scoring yourself</h4><p>Mark each test red, amber or green for your current estate. In our experience most organisations are green on one or two, usually retention or search, and red on governance.</p><p>A low score is not a reason to replace anything this year. It is a list of questions for your next vendor review and a way to decide where to pilot agents first: start where the governance answer is strongest, not where the demo looks best.</p><h4>2.4 A caution about demos</h4><p>Sharp made a fair point: leave any one of the six out and you still get an impressive demonstration. The differences show up at production scale, with real data volumes, real access rules and a real bill. Ask to see the audit trail and the cost report, not only the investigation.</p><p><em>Ask Apto for: a short readiness review against the six tests.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=c477f4f616da" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Cribl’s new SIEM: what it changes for your renewal, and ten questions to ask first]]></title>
            <link>https://medium.com/@apto.solutions/cribls-new-siem-what-it-changes-for-your-renewal-and-ten-questions-to-ask-first-287931b7ed8d?source=rss-1adf5dd40347------2</link>
            <guid isPermaLink="false">https://medium.com/p/287931b7ed8d</guid>
            <category><![CDATA[criblcon]]></category>
            <category><![CDATA[cribl]]></category>
            <category><![CDATA[cloud]]></category>
            <category><![CDATA[siem]]></category>
            <dc:creator><![CDATA[Apto Solutions]]></dc:creator>
            <pubDate>Mon, 16 Nov 2026 10:31:01 GMT</pubDate>
            <atom:updated>2026-11-16T10:31:01.773Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*CvcwUfqXLqsP0dNzXXxoyA.png" /></figure><p><em>What Cribl Detect means for your next SIEM renewal, and the due diligence a UK organisation should do before adopting a cloud-only SIEM</em></p><h3>1. Cribl is now a SIEM vendor. What that changes for your next SIEM renewal</h3><p><a href="https://www.youtube.com/watch?v=WnVfUH7xGm4&amp;t=372s"><strong><em>WATCH IN THE CRIBLCON 2026 KEYNOTE</em></strong></a></p><p><strong>3:37 Product expansion list</strong></p><p><strong>6:12 What customers mean by SIEM-less</strong></p><p><strong>1:07:22 Detect introduced</strong></p><p><strong>1:11:32 Pricing statement</strong></p><p>For eight years Cribl has been the neutral layer that sits in front of your SIEM, trimming and routing data so the bill stays tolerable. At CriblCon on 29 September it stopped being neutral. <a href="https://www.youtube.com/watch?v=WnVfUH7xGm4&amp;t=4042s">Cribl Detect</a> is a SIEM, and it is available now to Cribl.Cloud customers.</p><p>If a SIEM is your system of record for security, this does not mean you should move. It does mean your next renewal conversation has a third option that did not exist last month.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*RoIAhgAha0oLKgWt.png" /></figure><p><em>Figure 1 — This figure shows the run-alongside option: your SIEM keeps the data worth indexing while Cribl Detect covers the data that was too costly to index.</em></p><h4>1.1 What was announced</h4><p>Cribl’s chief executive, Clint Sharp, opened by saying <a href="https://www.youtube.com/watch?v=WnVfUH7xGm4&amp;t=217s">this was not last year’s presentation</a>: the company is expanding into SIEM, application monitoring and AI-driven security operations on one shared platform. He was candid about the reason. Customers keep telling him they want to go <a href="https://www.youtube.com/watch?v=WnVfUH7xGm4&amp;t=372s">“SIEM-less”</a>. What they mean, he said, is that they no longer want to pour everything into one vendor’s repository and be handed a larger bill a month before renewal.</p><p>Detect runs detections on data as it arrives in Cribl’s own store, before it is written to storage. Separately, it can search and investigate data where it already sits, including in other SIEMs; that is search, not detection. Cribl’s September release added the ability to search archived Splunk data in your own object storage without rehydrating it. It ships with a large library of pre-built detections, maps coverage to the MITRE ATT&amp;CK framework, and accepts Sigma and Splunk rules through what Cribl calls assisted translation.</p><h4>1.2 The commercial claim</h4><p>Cribl says Detect costs up to 50% less than other SIEMs. On stage the message was that <a href="https://www.youtube.com/watch?v=WnVfUH7xGm4&amp;t=4292s">the application itself carries no separate charge</a>; you pay for the infrastructure that ingests, stores and queries the data. Running detections adds to the capacity that infrastructure needs, so no separate charge does not mean no extra cost. That is a different meter from licence-by-ingest, and it deserves to be modelled against your own volumes, not taken on trust. The 50% figure is Cribl’s own and no methodology has been published.</p><h4>1.3 Three options at renewal</h4><p>Stay. If your detections are mature, your analysts are productive and your contract is fair, nothing here forces a change.</p><p>Run alongside. This is where most organisations will start. Almost every large SIEM estate has data it chose not to index because it was too expensive: DNS, flow logs, verbose cloud audit trails. That data can be sent to Cribl’s own store and detected on there, while the SIEM keeps doing what it does well. Data left in low-cost object storage can be searched and investigated, but confirm with Cribl whether detections can run on it, and model the cost of the store that detection needs. Independent analysts expect this to be the realistic first use.</p><p>Move workload by workload. Possible, but be clear about the effort. A practitioner quoted by TechTarget put it well: assisted translation means migration is “still a real project, not a switch”. Every detection needs validating and tuning on the other side.</p><h4>1.4 What to do before the renewal</h4><p>Go into the conversation knowing three numbers: how much data you ingest today, how much you discard or exclude because of cost, and how many of your detections actually fire. The second number is your negotiating position. The third tells you how large a migration would really be.</p><p>Two cautions. Detect is cloud-only today, which matters for regulated UK organisations and is the subject of a separate article. And Cribl now competes with vendors it has always partnered with, so watch how those relationships develop.</p><p>If your renewal falls in the next twelve months, we can help you size the run-alongside option before you sit down with either vendor.</p><p>Since the keynote: Clint Sharp’s own <a href="https://cribl.io/blog/how-i-think-about-detection-in-the-pipeline/">blog on detection in the pipeline</a> (3 September 2026) is candid that matching single events in the stream is fast but noisy, and that alerts a SOC can trust still need correlation against stored history. Two community apps on GitHub address Splunk migration directly: an <a href="https://github.com/Cribl-Community/cc-visicore-spl-to-kql">SPL-to-KQL translator</a> from VisiCore, and a <a href="https://github.com/Cribl-Community/cc-ufe-migrator">Universal Forwarder migrator</a>that converts inputs.conf settings into Cribl Edge sources.</p><h3>2. Cloud-only SIEM and UK data sovereignty: ten questions before you adopt Cribl Detect</h3><p><a href="https://www.youtube.com/watch?v=WnVfUH7xGm4"><strong><em>WATCH IN THE CRIBLCON 2026 KEYNOTE</em></strong></a></p><p><strong>1:08:21 Enabling Detect in Cribl.Cloud</strong></p><p><strong>21:20 Outpost for nodes behind the firewall</strong></p><p>Cribl Detect is an interesting product and we expect many of our customers to try it. It is also weeks old, available only in Cribl’s cloud, and sold on claims that have not yet been independently tested. This article sets out the ten questions we would want answered before putting security detection for a UK organisation on it.</p><p>None of these is a reason not to proceed. They are the due diligence a regulator, auditor or board would expect to see.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*R8i-MIT0ZYb7Oeqq.png" /></figure><p><em>Figure 2 — This figure shows where data sits and moves with Cribl Detect, and which of the ten questions applies at each boundary.</em></p><h4>2.1 Where it runs</h4><ol><li>Which cloud region will hold and process our data? Detect is <a href="https://www.youtube.com/watch?v=WnVfUH7xGm4&amp;t=4111s">switched on inside a Cribl.Cloud workspace</a>. Ask which regions are available, whether a UK region is among them, and whether any processing, including AI inference, happens elsewhere.</li><li>Is there a self-managed option? Not today. Press coverage reports that an on-premises version is under consideration for regulated industries. If your policy requires security tooling inside your own boundary, get the roadmap in writing with a date.</li><li>What stays on our side? Cribl’s Outpost feature lets you <a href="https://www.youtube.com/watch?v=WnVfUH7xGm4&amp;t=1280s">manage collection nodes behind your firewall</a> from a leader in your own environment. Detect can also search data held in your own storage, but search is not detection: plan on any data you want detections to run on being held in Cribl.Cloud. Map exactly which data leaves your estate and which is only queried in place.</li></ol><h4>2.2 Where the AI runs</h4><ol><li>Which AI models see our security data, and where are they hosted?Detect’s triage and investigation features use AI models, routed by Cribl. Ask which providers are involved, what data is sent to them, whether it is retained, and whether you can restrict the list.</li><li>Can we use the product with AI features off? Some organisations will want detection and search first and AI later, once their own AI policy is settled.</li></ol><h4>2.3 Whether it works for you</h4><ol><li>How much of our existing detection content will translate? Cribl offers assisted translation for Splunk and Sigma rules. A practitioner quoted by TechTarget noted that this means migration is “still a real project, not a switch”. Ask for a trial translation of twenty of your own rules and count how many work unchanged. Ask too how the rules will be extracted from your current SIEM; content already held as Sigma or as detection as code is likely to move most easily.</li><li>Can detections run on data held outside Cribl’s own store? The launch material presents working across data wherever it sits as a headline feature, but searching data in place and running detections on it are different things. Ask which stores detections can run against today, and get dates for the rest. Analysts at Futurum have asked how well detection holds up when one detection must correlate identity, cloud and endpoint data held in different places. Test a multi-stage scenario, not a single-source one.</li><li>Which features are live today? The written announcement lists AI-driven triage as part of Detect. On stage it was described as coming. Ask for a feature-by-feature availability list that covers AI triage, risk-based alerting, user behaviour analytics and detection as code.</li></ol><h4>2.4 What it costs and who you depend on</h4><ol><li>What is the 50% saving measured against? Cribl says Detect costs up to 50% less than other SIEMs and is priced on infrastructure, not on search volume. No methodology has been published. Ask how much extra capacity running detections adds to the infrastructure you pay for. Model it on your own volumes, retention and query patterns, including the growth you expect from AI agents.</li><li>What happens to our existing vendor relationships? Cribl now competes with SIEM vendors whose products it feeds. If you rely on Cribl in front of Splunk, Microsoft Sentinel or another platform, ask both sides how the integration will be supported over the life of your contract.</li></ol><h4>2.5 How to use this list</h4><p>Send it to your Cribl account team before a proof of concept, and treat the written answers as part of the evaluation. Where the answer is “not yet”, that is useful: it tells you which workloads to start with and which to hold back.</p><p>This article is not legal advice. Your data protection officer and, where relevant, your sector regulator’s guidance should determine what is acceptable for your organisation.</p><p>Apto Comment: Try our ten-question review and a scoped proof of concept on non-sensitive data.</p><p><em>Since the keynote: On question 7, Cribl’s chief executive has made the same point himself. His </em><a href="https://cribl.io/blog/how-i-think-about-detection-in-the-pipeline/"><em>3 September blog</em></a><em> says reliable alerts need cross-source, time-windowed correlation and full-fidelity history, and suggests asking any vendor to demonstrate both. On question 2, the </em><a href="https://cribl.io/blog/a-new-way-to-siem-starts-with-a-new-foundation/"><em>launch blog from Cribl’s head of security products</em></a><em> does not state deployment options or regions.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=287931b7ed8d" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The Agentic SOC Has a New Rule: Keep It on a Leash]]></title>
            <link>https://medium.com/@apto.solutions/the-agentic-soc-has-a-new-rule-keep-it-on-a-leash-709e37583157?source=rss-1adf5dd40347------2</link>
            <guid isPermaLink="false">https://medium.com/p/709e37583157</guid>
            <category><![CDATA[soc]]></category>
            <category><![CDATA[operations]]></category>
            <category><![CDATA[security]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[agentic-soc]]></category>
            <dc:creator><![CDATA[Apto Solutions]]></dc:creator>
            <pubDate>Thu, 12 Nov 2026 10:31:01 GMT</pubDate>
            <atom:updated>2026-11-12T10:31:01.841Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*JKR8ROJLh7tOgqNL8tAusg.png" /></figure><p>For the last two years, the pitch for AI in security operations has been speed. Let agents triage alerts, correlate signals and act faster than any human analyst could.</p><p>This month, the industry’s biggest vendors sent a different message: speed without control isn’t a feature, it’s a liability. Cisco’s post-.conf26 Splunk announcements, NVIDIA’s new agent-safety tooling and a run of reported agent failures all point the same way. 2026 is the year “trusted” quietly replaced “autonomous” as the word security vendors want attached to their AI.</p><h3>Constrained, not autonomous</h3><p>Splunk’s approach, detailed in <a href="https://www.techtarget.com/it-infrastructure/news/366650495/Splunk-security-exec-charts-course-for-responsible-AI">TechTarget’s .conf26 coverage</a>, is narrow by design. Rather than giving agents open-ended authority to investigate and remediate, the architecture leans on tightly scoped permissions and contextual guardrails. Agents get just enough latitude to be useful, and no more.</p><p>That is a deliberate contrast to the “autonomous SOC” framing that dominated vendor marketing a year or two ago. It reflects a maturing, if belated, recognition: an agent with broad access to logs, identities and infrastructure is itself an attack surface.</p><h3>On-prem and in-silicon: control moves closer to the metal</h3><p>Two parallel moves reinforce the point. Cisco’s new <a href="https://www.channelinsider.com/security/tools-and-platforms/cisco-splunk-ai-nvidia-on-premises/">on-premises Splunk AI option</a>, built with NVIDIA, gives data-sovereignty-conscious customers a way to run agentic workloads without sending telemetry to a third-party cloud. N.B. you do not need NVIDIA hardware, there is a software Tier only for Cisco AI Pod for Splunk.</p><p>Separately, NVIDIA’s <a href="https://developer.nvidia.com/blog/nvidia-open-agent-safety-platform-a-reference-for-continuous-in-silicon-agent-monitoring/">Open Agent Safety Platform</a> pushes monitoring down to the silicon layer. That means continuous, hardware-level oversight of what an agent is actually doing, rather than trusting its self-reported logs.</p><p>Together, these signal that control is becoming an infrastructure decision, not just a policy one.</p><h3>The incident that made it real</h3><p>None of this guardrail-building happens in a vacuum. <a href="https://thenewstack.io/anthropic-claude-agent-security/">The New Stack’s reporting</a> on Anthropic’s Claude agent failures has done more to concentrate minds than any vendor roadmap slide.</p><p>It is a visible reminder that agent observability isn’t an abstract best practice. It is the difference between catching a misbehaving agent in minutes and discovering the damage after the fact. Expect it to be cited in every “why agent observability matters” deck for the rest of the year.</p><h3>What MSSPs and buyers are actually doing about it</h3><p>On the ground, this is showing up as a market segment in its own right. <a href="https://seceon.com/ai-soc-platforms-for-mssps-compared-in-2026/">Comparison pieces</a> now evaluate AI SOC platforms on their oversight and audit capabilities, not just detection accuracy.</p><p>Cisco Secure Access also quietly shipped a three-click, <a href="https://blogs.cisco.com/security/security-insights-a-threat-first-view">threat-first investigation workflow</a>. It is a small UX change, but one aimed squarely at keeping a human in the loop without slowing the analyst down.</p><p>The common thread: buyers aren’t asking vendors to prove their agents are smart. They’re asking them to prove their agents are accountable.</p><h3>Takeaway</h3><p>The agentic SOC isn’t being abandoned. It’s being disciplined. The vendors making the most noise this month aren’t promising the most autonomy. They’re promising the tightest leash, the clearest audit trail, and the option to keep it all on-premises if you don’t trust the cloud with it.</p><p>For security teams evaluating this next generation of tooling, the right question in Q4 2026 isn’t “what can the agent do?” It’s “what can’t it do, and who’s watching when it tries?”</p><h3>Putting the leash on your own Splunk</h3><p>The same question applies the moment you let an AI assistant query your own Splunk. Splunk’s MCP Server makes that close to plug-and-play, which is exactly the risk: it is easy to switch on and hard to switch on well.</p><p>The work is in the decisions around it. Who can ask what, of which data, through which model, at what cost to the platform and your licence, and with what audit trail afterwards.</p><p>That is what Apto’s Splunk MCP Server Enablement service covers: a scoped role instead of a shared admin account, native guardrails, monitoring that ties usage back to your licence, and an owner for all of it under Apto Operate.</p><h4><a href="https://www.aptosolutions.co.uk/splunk/#splunkmcp">Find out more about Splunk MCP Server Enablement →</a></h4><h4>Sources</h4><ul><li><a href="https://newsroom.cisco.com/c/r/newsroom/en/us/a/y2026/m09/cisco-delivers-trusted-ai-at-scale-through-new-splunk-advancements.html">Cisco Delivers Trusted AI at Scale Through New Splunk Advancements</a>, Cisco Newsroom, 15 Sep 2026</li><li><a href="https://www.techtarget.com/it-infrastructure/news/366650495/Splunk-security-exec-charts-course-for-responsible-AI">Splunk’s path to autonomous AI starts with tightly constrained agents</a>, TechTarget, 15 Sep 2026</li><li><a href="https://www.channelinsider.com/security/tools-and-platforms/cisco-splunk-ai-nvidia-on-premises/">Cisco Brings Splunk AI On-Premises with NVIDIA</a>, Channel Insider, 17 Sep 2026</li><li><a href="https://www.techtarget.com/it-infrastructure/opinion/Splunk-conf26-Context-control-for-Ciscos-AI-agents-at-scale">Splunk .conf26: Context, control for Cisco’s AI agents at scale</a>, TechTarget, 17 Sep 2026</li><li><a href="https://developer.nvidia.com/blog/nvidia-open-agent-safety-platform-a-reference-for-continuous-in-silicon-agent-monitoring/">NVIDIA Open Agent Safety Platform: A Reference for Continuous In-Silicon Agent Monitoring</a>, NVIDIA Developer, 28 Sep 2026</li><li><a href="https://thenewstack.io/anthropic-claude-agent-security/">Anthropic’s Claude failures have made agent observability a security priority</a>, The New Stack, 2 Sep 2026</li><li><a href="https://blogs.cisco.com/security/security-insights-a-threat-first-view">Security Insights: A Threat-First View for the Platform That Enforces Access</a>, Cisco Blogs</li><li><a href="https://seceon.com/ai-soc-platforms-for-mssps-compared-in-2026/">AI SOC Platforms for MSSPs Compared in 2026</a>, Seceon via Security Boulevard, 30 Sep 2026</li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=709e37583157" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The Enterprise Just Hired Its First AI Workforce — Now It Has to Manage It]]></title>
            <link>https://medium.com/@apto.solutions/the-enterprise-just-hired-its-first-ai-workforce-now-it-has-to-manage-it-2967520db1c3?source=rss-1adf5dd40347------2</link>
            <guid isPermaLink="false">https://medium.com/p/2967520db1c3</guid>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[splunk]]></category>
            <category><![CDATA[conf26]]></category>
            <dc:creator><![CDATA[Apto Solutions]]></dc:creator>
            <pubDate>Mon, 09 Nov 2026 12:25:52 GMT</pubDate>
            <atom:updated>2026-11-09T12:25:52.065Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/768/1*BHaKIs9gPEZa9WbXkiBgsQ.png" /></figure><h3>Splunk .conf 26 Roundup!</h3><p>Every big tech conference claims a turning point. Most don’t have one. Splunk’s .conf26 in Denver this week past, might be the exception, and not because of a flashy new dashboard. Cisco and Splunk spent the week making a quieter, more consequential argument: AI agents have stopped being features and started being employees. And once you treat something like an employee, you have to start asking employee questions. Can you trust it? What does it cost? Who’s watching it when things go wrong?</p><p>We’ll run some future blogs on digesting this, but here is the first pass, of what’s hot off the press and some initial industry comments.</p><p>That thread runs through everything announced this week. Splunk’s own framing piece, from platform chief Mangesh Pimpalkhare, puts it bluntly: “The agentic era is no longer a promise on the horizon. It’s running in production, at scale, in the environments you depend on every day.” It’s worth following whether or not Splunk is your platform of choice, because the questions it raises aren’t Splunk questions. They’re 2027 budget questions.</p><p><a href="https://www.splunk.com/en_us/blog/leadership/digital-resilience-in-the-agentic-era-at-splunk-conf26.html">Splunk: Digital resilience in the agentic era</a></p><h4>Start with the keynotes</h4><p>Both broadcast keynotes are available on demand: the Welcome Keynote (1:41:25, Jeetu Patel and Kamal Hathi) and the Product Keynote (1:59:56). If you only have time for one clip, find Jeetu’s numbers. AI agents overtook humans in token consumption back in February 2026, and seven months later they’re burning five times as much. Roughly 60% of global AI compute now goes to inference rather than training. Then comes the line that carried the whole event: “agents are like teenagers — intelligent, fearless, and notorious for poor judgment.”</p><p>It’s a good line because it’s true, and because it does real work. It reframes every announcement that followed from “look what AI can do” to “here’s how we keep it supervised.”</p><p>Zeus Kerravala’s keynote analysis for SiliconANGLE reached the same conclusion from a different direction, and his title says it: trust becomes the product. His five takeaways are worth having to hand. Agents are workforce, not tooling. <em>Observability and security are merging</em> (we’ve been saying this for a while), because non-deterministic systems fail in ways that look exactly like breaches. Tokenomics turns AI from a science project into a P&amp;L line. Data economics have to change first. And agents earn trust the way employees do, on a trust-but-verify basis. Splunk’s Vikram Chatterjee supplied the quieter line of the week: “not everything needs a model in the loop.”</p><p><a href="https://conf.splunk.com/global-broadcast-replays.html">Watch the replays</a></p><p><a href="https://siliconangle.com/2026/09/16/trust-becomes-the-product-five-takeaways-from-the-splunk-conf26-keynotes/">SiliconANGLE: Trust becomes the product — five takeaways from the keynotes</a></p><h4>Agentic security operations grow up</h4><p>The headline product news is an expanded Agentic SOC Workforce: purpose-built agents across four lanes (Detection and Security Engineering, Threat Hunting, Investigation and Response, and Governance and Policy), now rolled into Enterprise Security Essentials and Premier. This isn’t a chatbot bolted onto a SIEM. It’s agents doing discrete SOC jobs end to end, with humans supervising rather than executing.</p><p>Read it plainly and it’s Splunk admitting the SOC hiring problem isn’t solvable by hiring. But the interesting agent is the least glamorous one. The Governance and Policy agent exists specifically to keep autonomy “grounded in SOPs, approvals, and auditability,” which tells you exactly where the buyer objection sits. It was never “can it detect threats.” It’s “can I audit what it did.”</p><p>The customer stories made that concrete. Asurion described a “trust ladder,” where agents earn expanded autonomy incrementally as they prove reliable, rather than being handed the keys on day one. Constellation Energy brought the number everyone will quote: identity-incident mean time to resolution fell from roughly 20 minutes to 39 seconds once they trusted automation to act rather than merely alert. That isn’t an efficiency gain. That’s a different category of response.</p><p><a href="https://www.splunk.com/en_us/blog/security/evolving-to-autonomous-defense-new-agentic-soc-capabilities.html">Read more: Evolving to Autonomous Defense</a></p><h4>The money question: what does an agent actually cost?</h4><p>This is the part that should worry, or at least interest, every budget owner. Agents are expensive, and the expense is invisible until the invoice lands.</p><p>Splunk’s answer is Agent Observability, now generally available and built on the Galileo acquisition. The standout feature is “Tokenomics”: a single view where finance and engineering can see which agents, users, and models are driving cost, and whether that spend is producing anything. Under the hood, small “Luna” models grade 100% of agent traffic cheaply, rather than paying a frontier model to judge a sample.</p><p>The real insight is the question shift, from “how much are we spending” to “which agents are wasting it.” Verizon’s contribution was the sharpest of the week: the cheapest agent isn’t always the most efficient one. A slow, chatty, imprecise agent can cost more in tokens than a pricier one that gets it right in fewer calls. Eighteen months ago nobody was asking for a FinOps layer for AI agents. By this time next year, after the first surprise bill, everyone will be.</p><p><a href="https://www.splunk.com/en_us/blog/observability/observe-evaluate-and-guardrail-your-ai-agents-with-agent-observability.html">Read more: Agent Observability</a></p><h4>Small models, not frontier models</h4><p>The Luna graders are part of a wider bet. Mike Vizard at DevOps.com reports that Splunk is preparing its second open-source LLM, this one trained specifically to reason across log data, headed for Hugging Face under an open licence. It complements the metrics-focused Cisco Time Series Model 1.0, and the logic is simple: general-purpose models are poor at raw numbers and worse at log volume, so build small ones that aren’t. Raja Mukhopadhyay, Splunk’s VP of observability cloud, says the log model will “make it possible for DevOps teams to observe AI applications and agents at scale.”</p><p>Pimpalkhare framed the strategy for SiliconANGLE without hedging: “We are not reinventing frontier models. Our goal is to complement those frontier models and develop the next level of domain-specific models.” Alongside it comes a new Universal Collector, built on OpenTelemetry to gather logs, metrics and traces through a single pipeline, with beta expected in 2027. The pitch is one collector feeding DevOps, IT operations and security at once, which is the plumbing that Kerravala’s “observability and security are merging” point depends on.</p><p><a href="https://devops.com/splunk-preps-second-open-source-llm-for-telemetry-data/">DevOps.com: Splunk preps second open-source LLM for telemetry data</a></p><h4>Splunk dismantles its own pricing model</h4><p>Two big moves landed in one post. First, Federated Search now reaches AWS CloudWatch Lake and Databricks under the Cisco Data Fabric banner, with SiliconANGLE’s deeper dive adding Snowflake and cloud object stores to the list, so you can analyse data where it lives instead of paying to ingest and duplicate it. Second, and more significant, Activity-Based Pricing launches this Fall: a dual-meter model weighting search and ingest equally, instead of penalising ingest volume.</p><p>Pimpalkhare’s line on why is refreshingly direct: “If you start duplicating data movement and all the processing needed, that is what drives up costs.” The stated goal, per Kerravala, is “capacity up 10x, bill flat”, and there’s an early proof point in the Autodesk story Network World picked up, with 28% lower ingest costs and a 78% saving from moving a single search workload.</p><p>It’s hard to overstate what this changes. Ingest-based pricing made Splunk expensive (our clients top bugbear) and shaped a decade of “what do we even bother logging” decisions. If agents are about to query constantly in the background, ingest-weighted pricing becomes a tax on curiosity, and Splunk knows it. Worth watching closely whether this lowers bills or simply moves where the meter sits. Either way, the model that built the company is being retired on purpose. We will get our Splunk contacts to reveal more watch this space.</p><p><a href="https://www.splunk.com/en_us/blog/platform/the-agentic-era-demands-a-new-data-foundation-how-cisco-is-delivering-it-with-splunk.html">Read more: The Agentic Era Demands a New Data Foundation</a></p><p><a href="https://siliconangle.com/2026/09/15/splunk-rebuilds-its-platform-around-agents/">SiliconANGLE: Splunk rebuilds its platform around agents</a></p><h4>Sovereignty gets a product line</h4><p>For regulated industries and anyone with data-residency constraints, the Cisco AI POD for Splunk is now generally available. It’s Splunk’s first on-premises accelerated-computing offering with NVIDIA: pre-validated, air-gap capable, Kubernetes-native, and running Splunk AI Assistant and Agent Launchpad locally. Model choice includes Gemma and GPT-OSS 20B today, with Nemotron coming.</p><p>It’s a direct answer to the question every bank, government agency, and healthcare provider has been asking: how do we get the benefits of agentic AI without sending sensitive data through someone else’s cloud model. It’s also a tell. If Splunk needed to build a whole on-prem SKU to get regulated customers to adopt agentic AI at all, that tells you how large the sovereignty objection really is right now. Expect this to matter most for public sector and financial services, but it signals a broader shift: AI capability is being decoupled from “must live in the cloud.”</p><p><a href="https://www.splunk.com/en_us/blog/partners/bringing-the-power-of-ai-to-where-your-data-lives-a-new-milestone-with-cisco-and-nvidia.html">Read more: Cisco AI POD for Splunk</a></p><h4>AWS and Splunk call it a category</h4><p>The expanded AWS partnership was positioned as “beyond a single integration”: adjustable autonomy (from recommend through to execute), policy-driven governance, and deep analyst oversight built into Splunk’s Agentic SOC on AWS infrastructure. It was light on specifics and heavy on positioning, which is itself the point. Two large vendors jointly telling the market that agentic security is a co-owned category matters more as a competitive signal than as anything you can buy today.</p><p><a href="https://www.splunk.com/en_us/blog/partners/splunk-and-aws-building-the-agentic-enterprise-together.html">Read more: Splunk and AWS</a></p><h4>Agent Launchpad: the most useful thing announced</h4><p>Quietly, this may be the most practically valuable announcement of the batch. Agent Launchpad is a no-code agent builder inside Splunk itself. Agents trigger directly from alerts and searches, inherit existing RBAC, and keep a full run-history audit trail. It removes the dependency on a data scientist: the analyst who already knows the runbook builds the agent.</p><p>One alpha customer, a SOC analyst at an ISP, reports their custom triage agent matches human severity classification around 80% of the time. That figure is honest in a way most vendor case studies aren’t. It doesn’t claim replacement. It claims triage support, and it tells you exactly where the human still belongs.</p><p><a href="https://www.splunk.com/en_us/blog/artificial-intelligence/introducing-splunk-agent-launchpad.html">Read more: Introducing Splunk Agent Launchpad</a></p><p>Rounding things out, Exposure Analytics picked up enhancements that shift the emphasis toward proactive risk prioritisation rather than reactive alert triage, consistent with the week’s wider theme of moving work upstream, before incidents happen.</p><h4>And for the network team</h4><p>Easy to miss among the agent headlines, but Network World’s write-up makes the case that infrastructure teams got real news too. A new Network Intelligence App for Splunk Observability pulls Cisco network topology, device health and event data into the same view as application telemetry, so an alert can be traced to the affected device without switching platforms or waiting for two teams to compare notes. ThousandEyes Network Insights adds synthetic testing across both the networks you control and the internet transit you don’t.</p><p>The motivation is agents again. Cisco’s Jeff Schultz noted that infrastructure is “now being strained at levels that it never was before” as distributed agents pile on traffic, and Kamal Hathi’s framing was that all of it “runs on a common set of data that can provide end-to-end visibility across network and application domains.” Agent Observability itself now stretches down to GPUs, vector databases and orchestration frameworks. If you run the platform, the agents are your workload now too.</p><p><a href="https://www.networkworld.com/article/4222273/highlights-from-splunk-conf26-and-what-the-news-means-to-network-engineers.html">Network World: What .conf26 means to network engineers</a></p><h4>What it means for buyers</h4><p>Strip away the branding and .conf26’s message is this: agentic AI in security operations is no longer a pilot-project conversation. It’s a procurement and governance conversation. The vendors in this space are now competing on trust, cost visibility, and deployment flexibility, not just detection accuracy.</p><p>Every announcement above is really answering the same question: how do you let an agent act without losing the ability to say why it acted, what it cost, and whether you’d have approved it. Capability was never the gap. Governance was.</p><p>If you’re evaluating any AI-driven security tooling this year, the questions worth asking aren’t “how smart is the agent.” They’re “how do we govern its autonomy, what does it cost us at scale, and where does our data actually have to sit.” Splunk isn’t the only vendor being forced to answer them. It’s just the one asking them loudest, in Denver, in front of everyone.</p><p>CRN’s channel read of the week draws the same five threads (the AI POD with NVIDIA, the Agentic SOC expansion, the AWS alliance, the platform overhaul, and the partner opportunity that sits underneath them) and lands on a point worth underlining. Every one of these announcements creates work that customers won’t do alone: sizing on-prem AI infrastructure, building and governing agents, rewiring data pipelines for federated search, and re-modelling costs under a new pricing meter. For partners, the agentic era looks less like a threat to managed services and more like a reason for them.</p><p><a href="https://www.crn.com/news/software/2026/5-key-takeaways-from-splunk-conf26">CRN: 5 key takeaways from Splunk .conf26</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=2967520db1c3" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Why Data Mapping Is the Foundation of Observability]]></title>
            <link>https://medium.com/@apto.solutions/why-data-mapping-is-the-foundation-of-observability-717f0e99e083?source=rss-1adf5dd40347------2</link>
            <guid isPermaLink="false">https://medium.com/p/717f0e99e083</guid>
            <category><![CDATA[data]]></category>
            <category><![CDATA[observability]]></category>
            <category><![CDATA[data-mapping]]></category>
            <dc:creator><![CDATA[Apto Solutions]]></dc:creator>
            <pubDate>Wed, 02 Sep 2026 16:03:17 GMT</pubDate>
            <atom:updated>2026-09-02T16:03:17.282Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*dScU0sKvE4snIizLbWSQXA.png" /></figure><h3>Understanding Your Data</h3><p>Every failed SIEM deployment and every underperforming observability platform has one thing in common: the organisation did not understand its data before it started building. They selected a platform, deployed agents, connected data sources, and started ingesting. Six months later, they had runaway costs, unreliable detections, and no clear picture of what they were actually collecting.</p><p>Data mapping — the disciplined process of discovering, inventorying, classifying, and documenting your telemetry data — is the unglamorous foundation that determines whether everything built on top of it works or fails.</p><h4>Data Classification: Three-Tier Model</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*mZZ8_V2O0nZK79oY.png" /></figure><p><em>Typically 10–20% of sources account for 60–80% of ingestion cost</em></p><h3>What Data Mapping Actually Means</h3><p><strong>Discovery</strong> identifies every source of telemetry in your environment — not just the obvious ones. Most organisations find 30–50% more data sources than expected, significant overlap, and critical services with no coverage at all.</p><p><strong>Flow mapping</strong> documents how data moves from source to destination. It reveals bottlenecks, single points of failure, unnecessary complexity, and duplication.</p><p><strong>Classification</strong> assigns a value and purpose to each data stream. Not all data is equally valuable, and treating it as such is the primary driver of cost overruns.</p><p><strong>Cost modelling</strong> quantifies the financial impact of each data stream. Typically, 10–20% of sources account for 60–80% of ingestion volume and cost.</p><h3>The Three-Tier Model</h3><p><strong>Tier 1 (High Value):</strong> Real-time ingestion into SIEM/observability platform at full fidelity. Security events, critical application logs, authentication/access logs.</p><p><strong>Tier 2 (Moderate Value):</strong> Aggregated or sampled before ingestion, or sent to warm storage. Infrastructure metrics, verbose application logs, network flow data.</p><p><strong>Tier 3 (Low Value/Compliance):</strong> Routed to cheap object storage for retention, not ingested into real-time platforms. Health checks, debug logs, routine audit trails.</p><h4>Data Mapping: Four-Step Process</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*c-1Id3SrCvs86jyD.png" /></figure><h3>How to Run a Data Mapping Exercise</h3><p><strong>Step 1: Inventory Your Sources</strong></p><p>Start with your SIEM and observability platforms — what is configured? Then expand to network infrastructure, cloud environments, and your application portfolio. Build a master inventory with consistent attributes.</p><p><strong>Step 2: Map the Flows</strong></p><p>For each source, trace the complete data flow from generation to final storage. Document every hop. Visualise the flows — a diagram immediately reveals architectural issues invisible in a flat inventory.</p><p><strong>Step 3: Classify by Value</strong></p><p>Assess each data stream across detection value, investigation value, and compliance value. Assign to Tier 1, 2, or 3.</p><p><strong>Step 4: Model the Costs</strong></p><p>Calculate the current cost of each data stream and the optimised cost based on your classification. The difference is your business case for implementing a telemetry pipeline.</p><p><em>A data map created once and filed away is a snapshot that decays immediately. The data map should be a living document, maintained as part of your operational model.</em></p><h3>Next Steps</h3><p>Ready to take action? Apto Solutions offers a range of entry-point engagements designed to give you clarity before commitment:</p><ul><li><strong>Free Assessment:</strong> A no-obligation conversation with one of our platform specialists to understand your current state and identify quick wins.</li><li><strong>SIEM Health Check:</strong> A structured review of your existing SIEM deployment covering architecture, detection coverage, data quality, and operational efficiency.</li><li><strong>Observability Maturity Assessment:</strong> A framework-driven evaluation of your monitoring and observability capabilities against industry best practice.</li><li><strong>Data Mapping and Discovery:</strong> An analysis of your telemetry data flows, identifying redundancy, gaps, and optimisation opportunities.</li></ul><p><a href="https://www.aptosolutions.co.uk/contact/"><strong>BOOK YOUR FREE ASSESSMENT</strong></a><strong> | Assessed. Architected. Operated</strong></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=717f0e99e083" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Splunk vs Microsoft Sentinel]]></title>
            <link>https://medium.com/@apto.solutions/splunk-vs-microsoft-sentinel-9aa68927b31f?source=rss-1adf5dd40347------2</link>
            <guid isPermaLink="false">https://medium.com/p/9aa68927b31f</guid>
            <category><![CDATA[siem]]></category>
            <category><![CDATA[splunk]]></category>
            <category><![CDATA[sentinel]]></category>
            <dc:creator><![CDATA[Apto Solutions]]></dc:creator>
            <pubDate>Mon, 08 Jun 2026 12:52:35 GMT</pubDate>
            <atom:updated>2026-06-08T12:52:35.284Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*sc9NitXN-t185DXFPLEZxg.png" /></figure><h3>Splunk vs Microsoft Sentinel: Which SIEM Is Right for Your Organisation?</h3><p>If you are evaluating SIEM platforms in 2026, the shortlist almost certainly includes Splunk Enterprise Security and Microsoft Sentinel. They dominate the UK enterprise market, but they represent fundamentally different architectural philosophies. Choosing between them is not about which is better — it is about which fits your environment, your team, and your operational model.</p><p>We operate both platforms for UK clients across financial services, healthcare, CNI, and mid-market enterprises. This comparison is drawn from that operational experience, not vendor marketing.</p><h4>Splunk vs Microsoft Sentinel: Key Differences</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*weOKY-nhLpWgD2mu.png" /></figure><h3>The Fundamental Difference</h3><p>Splunk is a data platform that happens to do security. Its architecture is built around SPL (Search Processing Language), a powerful, general-purpose query engine that can ingest and analyse virtually any machine data. Splunk Enterprise Security sits on top of this engine as a premium security application.</p><p>Microsoft Sentinel is a security platform built on Azure. It is tightly integrated with the Microsoft ecosystem — Defender XDR, Entra ID, Office 365, Intune — and uses KQL (Kusto Query Language) for search and analytics.</p><p>This distinction shapes everything: cost, flexibility, operational overhead, and the skills your team needs.</p><h3>Where Splunk Wins</h3><p><strong><em>Search flexibility.</em></strong> SPL is the most powerful query language in the SIEM market. If your security team has experienced Splunk engineers, they can build detection logic, data models, and investigations that simply are not possible in KQL. Risk-based alerting — Splunk’s framework for aggregating risk scores across entities rather than firing individual alerts — is particularly effective at reducing alert fatigue.</p><p><strong><em>Heterogeneous environments.</em></strong><em> </em>If your estate spans AWS, Azure, GCP, on-premises infrastructure, OT/SCADA networks, and a mix of Linux and Windows, Splunk’s breadth of data connectors and Technology Add-ons gives it a clear advantage.</p><p><strong><em>Observability convergence.</em></strong><em> </em>Splunk offers a parallel observability suite (Splunk Observability Cloud) that shares the same data platform. If you want to converge security and operational telemetry on a single platform, Splunk provides an architectural path that Sentinel does not.</p><p><strong><em>Deployment flexibility.</em></strong><em> </em>Splunk Cloud, Splunk Enterprise (self-managed), or hybrid. For organisations with strict data residency requirements — particularly in defence and government — the on-premises option matters.</p><h3>Where Sentinel Wins</h3><p><strong><em>Microsoft-heavy environments.</em></strong><em> </em>If your organisation runs Microsoft 365, Azure, and the Defender suite, Sentinel is the natural choice. Native data connectors provide rich, correlated security data with minimal configuration. Free ingestion of Microsoft telemetry dramatically reduces cost.</p><p><strong><em>Lower operational overhead.</em></strong><em> </em>Sentinel is fully managed PaaS on Azure. No infrastructure to manage, no indexer clusters to scale, no storage tiers to configure.</p><p><strong><em>Built-in UEBA.</em></strong><em> </em>Sentinel includes User and Entity Behaviour Analytics at no additional cost. In Splunk, behavioural analytics requires the separate UBA product with its own licence and infrastructure, although a premium version includes this and other add-ons now.</p><p><strong><em>Faster time-to-value.</em></strong><em> </em>Out-of-the-box analytics rules, hunting queries, and Logic Apps playbooks (native SOAR) mean Sentinel can deliver detection capability faster than Splunk, particularly for Microsoft-centric threat scenarios.</p><p><strong><em>Broader skill availability.</em></strong><em> </em>Azure and KQL skills are more widely available (and less expensive) than specialist Splunk engineers in the UK market.</p><h3>The Cost Question</h3><p>Raw licence cost is not the deciding factor — total cost of ownership is. Splunk’s licensing scales with ingest volume (GB/day). Without active data management, costs grow linearly with your data. The counter to this is pipeline optimisation: organisations that implement a telemetry pipeline between their data sources and Splunk typically achieve 30–60% ingestion reduction.</p><p>Sentinel’s pay-per-GB model looks attractive, especially with free Microsoft data ingestion. But Azure Log Analytics costs, Logic Apps consumption charges, and the operational cost of managing Sentinel at scale add up. Also consider where you store data long term, before you commence.</p><p>The honest answer: for Microsoft-heavy environments, Sentinel is usually cheaper. For heterogeneous estates, Splunk with pipeline optimisation is often more cost-effective when you factor in the breadth of visibility it provides.</p><h4>SIEM Platform Decision Framework</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*qnTVXFokbH8vsJLy.png" /></figure><h3>The Hybrid Option</h3><p>Increasingly, we see organisations running both — Sentinel for Microsoft-native telemetry and Splunk for everything else. A central telemetry pipeline (such as Cribl Stream) makes this practical by routing the right data to the right platform based on source, type, and cost profile.</p><h3>The Real Decision Factor</h3><p>Neither platform will deliver value without skilled operators running it day-to-day. Detection rules need tuning. Data quality needs managing. Alerts need triaging. The operational commitment is the same regardless of which platform you choose.</p><h4><em>What’s the most common failure?</em></h4><p><em>The most common failure mode we see is not a technology problem — it is an Operator gap. A managed SIEM service fills this gap and is often more impactful than the platform choice itself.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=9aa68927b31f" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Federated Security Analytics]]></title>
            <link>https://medium.com/@apto.solutions/federated-security-analytics-7f98bd5b1ea9?source=rss-1adf5dd40347------2</link>
            <guid isPermaLink="false">https://medium.com/p/7f98bd5b1ea9</guid>
            <category><![CDATA[analytics]]></category>
            <category><![CDATA[security]]></category>
            <category><![CDATA[federated]]></category>
            <category><![CDATA[vegas]]></category>
            <dc:creator><![CDATA[Apto Solutions]]></dc:creator>
            <pubDate>Wed, 20 May 2026 13:02:41 GMT</pubDate>
            <atom:updated>2026-05-20T13:02:41.203Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*7ttcjpYdpFpNct95rWFFzA.png" /></figure><h3>Why UK Enterprises Are Looking Beyond Centralised SIEM</h3><p>Security teams are caught in a familiar trap. The volume of telemetry worth analysing (cloud audit logs, identity events, EDR signals, SaaS application logs, network flow) keeps growing. The licence and storage budget for centralised SIEM does not. Something has to give, and increasingly it is the long-held assumption that all security data should live in one platform.</p><p>That assumption made sense when on-premises infrastructure produced predictable telemetry volumes and a single SIEM could realistically house it all. It makes much less sense when daily ingest growth outpaces budget growth and when significant categories of data (DNS, NetFlow, full EDR) are excluded from analysis simply because nobody can afford to ingest them.</p><h4>Centralised vs Federated</h4><p>In a centralised model, every event is shipped, parsed, indexed and stored in one analytics platform. Detections, hunts and investigations all run against that single store. The model is operationally simple, but it pushes every cost decision back to the ingestion question — what can we afford to keep?</p><p>Federated security analytics inverts the relationship. Rather than moving the data to the compute, the compute moves to the data. A federated query engine, such as Vega’s Security Analytics Mesh (SAM), reaches into your existing SIEMs, data lakes, object storage and XDR platforms, runs the query where the data already lives and returns the result — not the petabytes.</p><h4>But Don’t Pipelines Already Solve This?</h4><p>Telemetry pipelines such as Cribl Stream do meaningful work: routing, reducing and shaping data before it reaches the SIEM. Used well, they are the single biggest lever for SIEM cost control we see in client environments, and Apto deploys them constantly. What pipelines do not do is remove the underlying assumption that analytics happens in one place. They make the centralisation cheaper; they do not eliminate it. Federation tackles the architectural question that pipelines optimise around.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*AjJSre0REqNb55zK.png" /></figure><h4>Why This Matters for AI in the SOC</h4><p>Every analyst-augmenting AI system (investigation copilots, automated triage, hunt generation) depends on access to the underlying data. If half of your relevant telemetry is sitting in cold object storage because it was too expensive to index, the AI cannot reason about it. Federation makes the full data estate addressable, which turns out to be a precondition for AI to deliver in a SOC rather than just generate plausible-looking summaries of partial data.</p><h4>How It Plays Out in Practice</h4><p>Consider an investigation into a suspected credential compromise. The questions are not exotic: where did this account authenticate from, what cloud resources did it touch, what did the endpoint do, were there matching identities elsewhere. In a centralised model the answer depends entirely on what was ingested. In a federated model the same query reaches identity logs in Sentinel, cloud audit logs in a data lake, EDR data in a separate platform and historical archives in object storage — and returns one answer.</p><h4>About Vega</h4><p>For readers meeting Vega here for the first time: Vega is the security analytics company behind the Security Analytics Mesh (SAM). It was built on a simple bet — that the future of SIEM is fewer SIEMs, and that analytics should reach the data rather than the other way around.</p><p>SAM connects to Splunk, Microsoft Sentinel, CrowdStrike, Snowflake, Databricks, object storage and a growing list of identity, EDR and cloud sources. Detections, hunts and investigations run across all of them in place. Nothing is migrated. Nothing is re-indexed.</p><p>Vega has raised $185M from Accel, Cyberstarts, Redpoint and CRV, and is used today by Fortune 200 organisations, global banks and healthcare providers.</p><p>Vega calls this the Post-SIEM Era. For UK enterprises that have spent a decade negotiating ingest budgets line by line, the economics speak for themselves.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*I6AINQDM0gvsF2oh.png" /></figure><h4>How Apto Fits</h4><p>Federation is not a replacement for the operator gap; it changes the shape of it. Connectors need engineering. Detection content still needs writing, peer-reviewing and continuously validating. MITRE coverage still needs measuring. Egress and federated query costs need monitoring. None of this happens by accident.</p><p>This is where Apto fits. We already operate the platforms a federated mesh sits on top of — Splunk, Grafana, Microsoft Sentinel, Cribl, Databricks. Adding Vega as a federation layer means the operator already understands the underlying data, the licence economics, the schemas and the existing detection content. Our Operate Core service monitors SAM performance, source connector health and MITRE coverage daily; Operate Attach delivers the engineering work — detection-as-code pipelines, connector onboarding, KQL library development — in prioritised sprints.</p><p>For UK enterprises facing the same data-growth and budget-compression dynamics that drove this architectural shift, federation is becoming hard to ignore. The technology is ready. The harder question is who operates it day to day. We think we have a useful answer to that.</p><p><strong>Find Out More</strong></p><ul><li><a href="https://www.aptosolutions.co.uk/contact/"><strong><em>Book a free federated analytics assessment</em></strong></a></li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=7f98bd5b1ea9" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[What is a Managed SIEM Service?]]></title>
            <link>https://medium.com/@apto.solutions/what-is-a-managed-siem-service-0783c455c379?source=rss-1adf5dd40347------2</link>
            <guid isPermaLink="false">https://medium.com/p/0783c455c379</guid>
            <category><![CDATA[managed-services]]></category>
            <category><![CDATA[managed-siem]]></category>
            <dc:creator><![CDATA[Apto Solutions]]></dc:creator>
            <pubDate>Thu, 07 May 2026 09:31:01 GMT</pubDate>
            <atom:updated>2026-05-07T09:31:01.226Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*8-6TzpnpNyHJrmlb1ygoVA.png" /></figure><h3>What is a Managed SIEM Service?</h3><p>Your organisation invested six figures in a SIEM platform. You deployed it, connected your data sources, and built some dashboards. Six months later, the detection rules have not been updated, the alert queue is full of false positives nobody investigates, data quality has drifted, and the platform is consuming budget without delivering proportional security value.</p><p>This is not a technology failure. It is an operations failure. And it is the single most common pattern we see across UK enterprises running Splunk, Sentinel, and every other SIEM platform.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*cLyrwuO_r69CBIKP.png" /></figure><h3>The Operator Gap</h3><p>Every SIEM deployment has three roles that need to be filled:</p><p><strong><em>Users</em></strong> are the security analysts who consume SIEM outputs — investigating alerts, running searches, producing reports. Most organisations have these.</p><p><strong><em>Builders</em></strong> are the engineers who deploy and configure the platform — standing up infrastructure, connecting data sources, building initial detection rules. These are typically project resources who move on after deployment.</p><p><strong><em>Operators</em></strong><em> </em>are the people who run the platform day-to-day. They tune detection rules as the threat landscape evolves. They manage data quality as new sources are added and existing sources change format. They optimise costs as data volumes grow. They ensure the platform continuously improves rather than gradually degrading.</p><p>Most organisations have Users and Builders. Almost none have dedicated Operators. This is the Operator gap, and it is why the majority of SIEM deployments underperform within 12 months of going live.</p><h3>What a Managed SIEM Service Actually Includes</h3><p>A managed SIEM service fills the Operator gap by providing dedicated platform operations expertise on a continuous basis. It is not outsourced SOC (that is a different service). It is the operational layer that ensures your SIEM platform works effectively so that your internal team can focus on investigating real threats rather than wrestling with the tooling.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*Mtw57eXLuqoMC6Yg.png" /></figure><p><strong>Platform health and performance. </strong>Continuous monitoring of the SIEM infrastructure — search performance, indexing throughput, storage capacity, cluster health. Proactive intervention before performance degradation impacts analysts.</p><p><strong>Detection engineering. </strong>Ongoing development and tuning of detection rules, correlation searches, and alerting logic. This includes mapping to frameworks like MITRE ATT&amp;CK, responding to new threat intelligence, and adapting rules as your environment changes.</p><p><strong>Data quality management. </strong>Monitoring that all expected data sources are reporting, that data formats have not changed, that parsing and normalisation are working correctly, and that tagging standards are being followed.</p><p><strong>Cost optimisation.</strong> Active management of data volumes to prevent licence cost spiralling. This includes identifying and filtering low-value data, recommending aggregation strategies, and managing data retention policies.</p><h3>Managed SIEM vs Managed SOC</h3><p>A Managed SOC provides human analysts who monitor your alerts, triage incidents, and escalate threats. The focus is on the security workflow. A Managed SIEM focuses on the platform itself — ensuring it runs efficiently, detects accurately, and evolves continuously. It is the operations layer underneath the SOC.</p><p>You can have one without the other. Many organisations run their own SOC but use a managed SIEM service for platform operations because they lack the specialist skills to maintain the platform.</p><h3>When Does a Managed SIEM Service Make Sense?</h3><ul><li>Your internal security team is strong on analysis but lacks specialist platform operations skills</li><li>You cannot recruit or retain dedicated SIEM engineers</li><li>Your SIEM platform has been deployed but is not delivering expected detection coverage</li><li>Your SIEM costs are growing faster than your security budget</li><li>You want a co-managed model where your team handles triage while a partner handles platform ops</li></ul><h4>Next Steps</h4><p>Ready to take action? Apto Solutions offers a range of entry-point engagements designed to give you clarity before commitment:</p><ul><li><strong><em>Free Assessment:</em></strong> A no-obligation conversation with one of our platform specialists to understand your current state and identify quick wins.</li><li><strong><em>SIEM Health Check:</em></strong> A structured review of your existing SIEM deployment covering architecture, detection coverage, data quality, and operational efficiency.</li><li><strong><em>Observability Maturity Assessment:</em></strong> A framework-driven evaluation of your monitoring and observability capabilities against industry best practice.</li><li><strong><em>Data Mapping and Discovery:</em></strong> An analysis of your telemetry data flows, identifying redundancy, gaps, and optimisation opportunities.</li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=0783c455c379" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Managed SOC vs Managed SIEM]]></title>
            <link>https://medium.com/@apto.solutions/managed-soc-vs-managed-siem-9877bb098aed?source=rss-1adf5dd40347------2</link>
            <guid isPermaLink="false">https://medium.com/p/9877bb098aed</guid>
            <category><![CDATA[managed-services]]></category>
            <category><![CDATA[siem]]></category>
            <category><![CDATA[soc]]></category>
            <dc:creator><![CDATA[Apto Solutions]]></dc:creator>
            <pubDate>Thu, 30 Apr 2026 08:31:12 GMT</pubDate>
            <atom:updated>2026-04-30T08:31:12.187Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*cwugZ0yEKG-06LvxlWvG9Q.png" /></figure><h3>Managed SOC vs Managed SIEM: Understanding the Difference</h3><p>The terms managed SOC and managed SIEM are often used interchangeably in vendor marketing, but they describe fundamentally different services that solve different problems. Confusing them leads to buying the wrong service and leaving critical gaps in your security operations.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*7fS7J7gb90MDW50D.png" /></figure><h4>What a Managed SOC Provides</h4><p>A Managed SOC (Security Operations Centre) is a service that provides human security analysts who monitor your environment, triage alerts, investigate incidents, and coordinate response. The focus is on the security workflow — the process of detecting, analysing, and responding to threats.</p><p>A managed SOC typically provides 24/7 monitoring and alert triage, incident investigation and escalation, threat hunting and proactive detection, incident response coordination, and regular reporting on security posture.</p><h4>What a Managed SIEM Provides</h4><p>A Managed SIEM service focuses on the platform itself — the technology that collects, correlates, and presents security data. The focus is on ensuring the SIEM works effectively as a tool.</p><p>A managed SIEM service covers platform health monitoring, detection rule engineering and tuning, data source management and quality, cost optimisation and licence management, and continuous platform improvement.</p><h4>The Relationship Between Them</h4><p>A managed SIEM is a subset of the operational capability that a SOC requires. You can think of it as the platform layer underneath the security operations layer. A SOC needs a well-operated SIEM to be effective. But a well-operated SIEM does not automatically give you SOC capabilities.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*t_k0LcnO0wJs0KTV.png" /></figure><h4>Which Do You Need?</h4><p><strong><em>Managed SIEM if: </em></strong>you have an internal security team that handles alert triage and incident response, but your SIEM platform is underperforming. Your team has the security skills but lacks the platform engineering expertise to keep the SIEM running effectively.</p><p><strong><em>Managed SOC if:</em></strong><em> </em>you do not have an internal security team with the capacity for 24/7 monitoring, investigation, and response. You need both the platform operations and the security operations provided as a service.</p><p><strong><em>Both if:</em></strong><em> </em>you want a specialist partner managing the platform while another team (internal or external) handles the security workflow. This is common in organisations that are building internal SOC capability but need platform support.</p><p><strong><em>Be Aware: </em></strong>lots of external managed SOC providers claim to have Managed SIEM incorporated, like our first diagram. This is not true on many occasions, as they use their own SIEM or platform, and it’s a entirely different and separate service. Do your due diligence any true managed SIEM provider can explain the difference.</p><h3>What should I start with?</h3><p>Start with Managed SIEM if your primary problem is platform performance and detection quality. Add SOC capabilities when you need analyst coverage beyond what your internal team can provide.</p><h3>Next Steps</h3><p>Ready to take action? Apto Solutions offers a range of entry-point engagements designed to give you clarity before commitment:</p><ul><li><strong><em>Free Assessment: </em></strong>A no-obligation conversation with one of our platform specialists to understand your current state and identify quick wins.</li><li><strong><em>SIEM Health Check: </em></strong>A structured review of your existing SIEM deployment covering architecture, detection coverage, data quality, and operational efficiency.</li><li><strong><em>Observability Maturity Assessment:</em></strong> A framework-driven evaluation of your monitoring and observability capabilities against industry best practice.</li><li><strong><em>Data Mapping and Discovery:</em></strong> An analysis of your telemetry data flows, identifying redundancy, gaps, and optimisation opportunities.</li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=9877bb098aed" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[SIEM vs XDR in 2026: Why the Question Has Changed]]></title>
            <link>https://medium.com/@apto.solutions/siem-vs-xdr-in-2026-why-the-question-has-changed-8d6db0050df8?source=rss-1adf5dd40347------2</link>
            <guid isPermaLink="false">https://medium.com/p/8d6db0050df8</guid>
            <category><![CDATA[siem]]></category>
            <category><![CDATA[soc]]></category>
            <category><![CDATA[agentic]]></category>
            <category><![CDATA[xdr]]></category>
            <dc:creator><![CDATA[Apto Solutions]]></dc:creator>
            <pubDate>Thu, 16 Apr 2026 10:01:02 GMT</pubDate>
            <atom:updated>2026-04-16T10:01:02.061Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*VkKUl1qZH37u90UUz7IkbQ.png" /></figure><p>When we first published this <a href="https://www.aptosolutions.co.uk/blog/siem-vs-xdr/">blog</a> in 2024, the cybersecurity world was wrestling with a familiar question: should we invest in a SIEM, or move to XDR? A year on, the landscape has shifted so dramatically that the question itself feels outdated. The categories are converging, AI agents are rewriting the rules of security operations, and the vendors you know have made bold platform bets that blur every traditional boundary. Here’s what’s really happening — and what it means for your security strategy in 2026 and beyond.</p><h3>A Quick Recap: The 2024 Debate</h3><p>In 2024, the SIEM-vs-XDR conversation was largely about trade-offs. SIEM gave you broad log aggregation, compliance reporting, and deep historical visibility — but at the cost of alert fatigue, slow investigations, and high operational overhead. XDR promised tighter integration across endpoints, network, and cloud, with better correlation and faster response — but lacked the maturity for regulatory compliance and long-term data retention. Most organisations were running both in parallel, stitched together with SOAR playbooks and a lot of manual effort.</p><p>The catalyst for change came from the vendor side. Cisco’s Splunk acquisition raised fundamental questions about the future of standalone SIEM. Clients were asking us: does it still make sense to invest heavily in a SIEM when XDR platforms are absorbing more and more of the detection and response workflow? Our original answer was nuanced — both tools have a role, and the right choice depends on your organisation’s maturity, regulatory obligations, and existing investments.</p><p>That answer was correct then. But the world of siloed tools, console switching, and Tier-1 analysts drowning in false positives? It’s rapidly disappearing.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*DExrg9CaJbpLv42a.png" /></figure><h3>What Changed: The Great Convergence</h3><p>The biggest shift in 2024–26 hasn’t been one product beating another — it’s been the categories themselves merging. SIEM, XDR, and SOAR are collapsing into unified security platforms. Palo Alto Networks absorbed IBM QRadar SaaS customers and folded XDR, SOAR, ASM, and SIEM into Cortex XSIAM. Microsoft expanded Sentinel with native XDR through Defender, creating a single operational layer. Google evolved Chronicle into Google SecOps with integrated detection and response. CrowdStrike pushed Falcon Next-Gen SIEM to unify log management with its XDR capabilities.</p><p>And then there’s Cisco. Having completed the $28 billion Splunk acquisition, Cisco has taken a distinctive approach: rather than replacing one tool with another, they’re bridging them. At Cisco Live Amsterdam 2026, the SOC demonstration showcased Cisco XDR and Splunk Enterprise Security working as a closed-loop integration — XDR acting as the real-time triage engine, receiving telemetry and producing high-fidelity incident bundles at machine speed, while Splunk ES provides the deep analytics backend with custom OCSF detections and long-term data retention. The release of Splunk Enterprise Security 8.2 formalised this with two new packaging tiers: ES Essentials (SIEM plus AI Assistant) and ES Premier (adding SOAR, UBA, and threat intelligence management). Crucially, Federated Search now lets analysts query data across Splunk Cloud and Cisco’s Security Analytics and Logging without ingestion — reducing cost while keeping full investigative reach.</p><p>The modern SIEM market is projected to grow from $7.13 billion in 2024 to $13.55 billion by 2029 at a 13.7% CAGR — but what’s growing isn’t the SIEM your team deployed five years ago. What’s dying is legacy SIEM as passive log storage. What’s emerging is SIEM as an AI-powered security intelligence platform that has absorbed XDR and SOAR capabilities wholesale.</p><h3>The Agentic SOC: AI Takes the Analyst’s Seat</h3><p>Perhaps the most transformative development is the rise of agentic AI in the SOC. In February 2026, Palo Alto Networks launched Cortex AgentiX — a platform for building, deploying, and governing AI agents that autonomously investigate alerts, not by following fixed if-then playbooks, but by reasoning about novel scenarios and adapting investigation paths in real time. Early adopters are reporting up to a 98% reduction in mean time to respond (MTTR) with 75% less manual work.</p><p>This isn’t just about speed. The agentic approach fundamentally changes the SOC operating model. When an alert fires, the system acts like an experienced analyst: it asks follow-up questions, fetches missing context from across your estate, re-evaluates the signal, and either resolves the incident or escalates it with a fully enriched case file. Microsoft’s Security Copilot, CrowdStrike’s Charlotte AI, and Google’s Gemini in SecOps are all racing toward the same vision. Cisco is right in this race too: Splunk ES 8.2 introduced a Triage Agent, AI Playbook Authoring, and a Personalised Detection SPL Generator — capabilities that let the platform auto-generate and tune detection rules based on your specific environment rather than relying on generic rule libraries.</p><p>For mid-market organisations running lean security teams, this is a game-changer. You no longer need a twenty-person SOC to achieve enterprise-grade detection and response — provided you architect your platform correctly and pair it with the right managed service support.</p><p>We’re already seeing this with our own Operate clients. Organisations that previously needed three or four full-time analysts to manage separate SIEM and EDR consoles are now operating effectively with a smaller, more strategic team augmented by AI-driven triage. The analyst role isn’t going away — it’s being elevated from alert factory worker to security strategist.</p><h3>The Hidden Enabler: Security Data Pipelines</h3><p>One trend that the Software Analyst newsletter has rightly spotlighted is the rise of Security Data Pipeline Platforms (SDPPs). AI-driven detection only works when data is normalised, enriched, and routed intelligently across your stack. The Open Cybersecurity Schema Framework (OCSF), now under the Linux Foundation, is becoming the backbone of vendor-agnostic data standardisation — enabling federated search, cross-platform correlation, and portable detection rules regardless of which SIEM you run.</p><p>Vendors like Cribl, Tenzir, and Monad are thriving in this space. SentinelOne’s $225 million acquisition of Observo AI in late 2024 signalled that even the XDR-native players now see data pipeline control as strategic. Cisco’s own Federated Search capability reflects the same insight — letting analysts query firewall logs stored in Security Analytics and Logging directly from Splunk Cloud without costly re-ingestion. If SIEM was once about collecting logs, the modern platform battle is about controlling the data fabric that feeds every detection, investigation, and response workflow.</p><p>This matters because detection engineering itself is being transformed. As SACR’s research on the future of detection engineering highlights, the old model — either relying on noisy generic rule libraries or investing heavily in bespoke detection-as-code — is giving way to AI-driven detection generation that continuously tunes rules based on observed environment context. Detections are increasingly treated not just as logic, but as logic plus threat model, risk criteria, and investigation guidance — encoded like software and governed with the same rigour. The convergence of AI and data security means platforms must now defend systems that reason, interpret intent, and execute autonomously, requiring a fundamentally different security posture than the deterministic software models of the past.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*4FljHYuLrcA7FWbE.png" /></figure><h3>The Consolidation Risk You Can’t Ignore</h3><p>There is a flip side to convergence that deserves honest discussion: vendor lock-in. When your SIEM, XDR, SOAR, and now AI agent layer all come from a single vendor, switching costs become enormous. The organisations getting this right are the ones investing in open data standards — particularly OCSF — so their detection logic and data remain portable even as they adopt a primary platform vendor. Think of it as insurance: commit to a platform, but keep your data liberated.</p><h3>What This Means for Your Organisation</h3><p>If you’re still running a legacy on-premises SIEM and bolting on separate EDR/XDR, the strategic risk is real. You’re paying a “complexity tax” — duplicated data, manual correlation, and analyst time spent switching between consoles rather than investigating threats. The convergence trend means the gap between what your current stack can do and what a modern platform delivers is widening every quarter.</p><p><strong>Here’s our practical advice for 2026:</strong></p><p><strong><em>Stop thinking in categories.</em> </strong>Don’t ask “SIEM or XDR?” — ask “which unified platform best fits my data sources, compliance requirements, and existing ecosystem?” The answer will depend on whether you’re a Microsoft shop, a Palo Alto estate, a Cisco/Splunk environment, or running multi-vendor.</p><p><strong><em>Invest in your data layer. </em></strong>Before you migrate platforms, get your security data pipeline right. OCSF-normalised, well-routed telemetry makes every platform perform better and gives you portability if you need to switch.</p><p><strong><em>Pilot agentic capabilities.</em></strong> The AI SOC isn’t vapourware anymore — it’s shipping. Run a controlled pilot with your most common alert types and measure the MTTR improvement. The results will build the business case.</p><p><strong><em>Don’t go it alone. </em></strong>Platform convergence and AI adoption demand specialist knowledge. A managed service partner who understands both the legacy migration path and the new platform architectures can compress your timeline from quarters to weeks.</p><h3>Where Apto Solutions Fits</h3><p>At Apto Solutions, our Operate managed service is built for exactly this moment. We help organisations navigate the SIEM-to-platform transition — whether that’s migrating from legacy Splunk to Cortex XSIAM, optimising a Microsoft Sentinel deployment, or architecting a hybrid approach that keeps compliance intact while modernising detection and response. We handle the platform management, detection engineering, data pipeline optimisation, and operational overhead so your team can focus on strategic security outcomes rather than keeping the lights on. Our approach is vendor-aware but outcome-driven — we work with your existing investments, not against them.</p><p>The SIEM-vs-XDR debate served its purpose. In 2026, the real question is: how fast can you get to a unified, AI-augmented security operations platform — and do you have the right partner to get you there?</p><p>The security operations landscape in 2026 rewards organisations that think in platforms, invest in data quality, and embrace AI as a force multiplier rather than a threat to their SOC teams. The analysts who thrive will be the ones freed from repetitive triage to focus on threat hunting, purple teaming, and strategic risk reduction — the work that actually moves the needle.</p><p><strong>Ready to modernise your security operations?</strong></p><p><a href="http://aptosolutions.co.uk/contact"><strong>GET IN TOUCH WITH APTO SOLUTIONS TO DISCUSS YOUR PLATFORM STRATEGY</strong></a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=8d6db0050df8" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>