MDR is not going away, but the old model is. The old model was service-first: collect alerts, have analysts investigate them, escalate what matters, and report what happened. That still solves a real problem, but it is not where the market is going.

The future MDR has to become product-first, AI-native, and action-oriented. It cannot just sit on top of EDR alerts and add analyst labor. It has to become the SecOps control plane that connects detection, investigation, response, exposure reduction, and proof of outcome.

This is not just an “AI layer” on top of alerts. It requires a different architecture, a different relationship with the SIEM, and a different operating model for response.

The Service Business Has To Become A Product Company

Traditional MDR was built as a service business. The product was often the people: analysts, playbooks, escalations, customer calls, and reporting. That model can scale to a point, but it starts to break when the market expects software-speed investigation and response.

Next-gen MDR has to invert the model. The platform becomes the core product, and services become the trust, expertise, and exception-handling layer around it. The test is margin and labor leverage. If more customers create more exceptions, more tuning, and more analyst burden, the company is still a service business. If more customers create more reusable detections, known-good patterns, automation recipes, and response policies, the service is becoming a product. Analysts still matter, but they should supervise, improve, validate, and handle edge cases. They should not be the primary execution engine for every repetitive investigation.

That is a hard transition because many MDR providers are not software companies. They can script, tune, and operationalize, but building a real SecOps platform is a different muscle.

The Moat Is Not AI. It Is Encoded Operational Judgment.

The defensible part of next-gen MDR is not that it uses AI. Everyone will use AI. The defensible part is whether the provider has converted years of SOC judgment into productized decision systems: known-good behavior, investigation procedures, response playbooks, customer-specific rules of engagement, approval policies, and feedback loops that improve the platform over time.

The best MDR providers will not simply have analysts assisted by AI. They will have a compounding operating system where every resolved case can improve future triage, investigation, automation, and response across customers. The weak version is service history trapped in analyst memory and exception lists. The strong version is structured, versioned, measurable operational knowledge embedded into the platform.

And let’s not forget the context graph. The data structure that learns the customer environments. The intelligence layer that fuses together individual data points to deliver a high signal, low friction security outcome.

MDR And SIEM Become Symbiotic

It’s a mistake think (or build a company around) next-gen MDR replaces the SIEM.

The SIEM is there for a reason. It stores logs, supports compliance retention, enables historical search, runs detections, and gives the customer a system of record for security data. Many customers already have years of detection logic, parsers, dashboards, and workflows inside it.

But the SIEM should not be the only operating layer for MDR. If MDR depends entirely on the (customer’s) SIEM, it inherits the SIEM’s latency, schema problems, data opacity, cost model, and workflow limitations.

The future relationship is symbiotic. MDR should use the SIEM’s detection layer where it works, but also audit it, monitor it, and improve it. Are the right logs arriving? Are detections still firing correctly? Are parsers broken? Are detections duplicated, stale, or missing context? Which investigation outcomes should feed back into detection logic?

In that model, the SIEM becomes less of the SOC’s user interface and more of a data and detection layer inside a broader SecOps operating system. It’s also feasible to have the SIEM be the data backend for the AI SOC, but only if it’s under the uninhibited control of the MDR.

AI Changes The Architecture, Not Just The Interface

The visible change in the near future is that analysts and customers will interact with the system(s) through AI. They will ask questions, start investigations, summarize cases, generate response plans, even design their own dashboards, and check what changed.

But the harder shift is underneath. The backend has to be built so AI can interact with it safely and efficiently. That means the platform needs clean APIs, stable entity models, permission-aware access, auditable actions, reusable investigation workflows, and compact context that does not require dumping endless raw logs into a model. And last but not least, the backend has to be designed for AI workloads and in turn, for token efficiency and for the types of (probably very suboptimal) queries that the AIs will send down.

This is why an AI-native SecOps platform is not the same as a chatbot on top of a SIEM. The architecture has to support the way AI investigates, acts, remembers, and learns.

Detection Has To Move Into Automated Response

The current MDR model still assumes a human-centered response loop: a detection fires, an analyst investigates, and someone decides whether to isolate, disable, block, remove, patch, ticket, or escalate. That loop is too slow for many modern attacks.

The future MDR platform has to move toward detection-to-response automation, but not reckless automation. Customers will push back, and rightly so. They do not want an opaque AI system taking production-impacting actions on its own. The answer is to be precise about what “AI” means.

Not every workflow needs an LLM. Much of response should become deterministic rules, policy engines, and pre-approved playbooks. GenAI should be reserved for the places where it helps: summarization, investigation planning, evidence interpretation, query generation, and edge-case reasoning. For most alerts, the system should be predictable, auditable, and constrained.

The second answer is supervision. Some actions can be fully automated. Some should require analyst approval. Some should require customer approval. The platform has to know the difference between removing a malicious email, disabling a privileged identity, isolating a production server, and changing access policy for a sensitive application.

The third answer is guardrails. We are entering the era of validation loops for AI: agents and control layers that check what the system proposes, understand disruption risk, enforce policy boundaries, and monitor whether the action worked. Formal guardrail frameworks, policy-as-code, approval gates, rollback paths, evals, and audit trails become part of the SecOps architecture.

This is where risk becomes the control signal. A suspicious event on a low-value test asset should not trigger the same response as the same event on a privileged identity accessing sensitive data. Risk should influence not only the analyst queue, but also resource access flows: step-up authentication, zero trust enforcement, restricted data access, endpoint isolation, email removal, identity reset, or escalation to a human.

That is the important shift. Detection should not only create an investigation. It should drive protection and exposure reduction.

The Market Implication

Next-gen MDR, AI SOC, and the MSP/MSSP security platform may look like separate markets today. Architecturally, they are converging. The packaging changes by segment, but the platform direction is increasingly the same.

Enterprise buyers will usually want to retain ownership of security operations. They may use managed expertise, but they are more likely to buy an AI-native SecOps platform that compresses L1/L2 investigation, improves consistency, and accelerates response while their internal team keeps control of higher-risk decisions. There is an opportunity for AI SOC players to provide the new sec ops platform for these enterprises as the central command – AI first, SIEM symbiotic.

Mid-market buyers are the most interesting and the most difficult. They often have enough security maturity to know they need help, but not enough staff or process maturity to run a pure platform themselves. This is where productized MDR can become the managed SecOps control plane: part platform, part expert service, part automation layer. The provider has to help operate the loop, not just forward alerts. The lesson is not to bolt on every adjacent service. Exposure management, asset visibility, identity posture, vulnerability prioritization, and response orchestration all matter, but they have to reinforce the MDR operating loop. If the provider adds too much service scope without product leverage, it becomes broad, expensive, and hard to deliver. This is also where we need the vendors to educate the market. Mid-market companies need these capabilities whether they ask for them or not. If they are buying separate tools for exposure management, asset inventory, vulnerability management, identity posture, and reporting, it is reasonable to ask why the MDR provider should not help operate those workflows. The customer can still supervise decisions, set policy, and approve sensitive actions. But the MDR should increasingly coordinate the operating loop. Ideally MDR players will open up their platforms for end customers to co-deliver security results.

SMB buyers may never experience this as an AI SOC platform. They may experience it through an MSP or MSSP security platform that bundles endpoint, email, identity, compliance, patching, response, and reporting all within one simple, per user license. But underneath, the same direction applies: more product, more automation, more risk context, more proof.

The future MDR is not an AI layer on top of alerts. It is not a cheaper SOC. It is not a duplicate SIEM. And it is not a service desk with better automation. The future MDR is a productized SecOps control plane: symbiotic with the SIEM, directly integrated with the customer’s enforcement tools, risk-aware in how it prioritizes and responds, deterministic where safety matters, AI-native where reasoning helps, and measured by proof of risk reduction. The companies that win will be the ones that turn service experience into compounding product advantage. The companies that lose will still have analysts, alerts, dashboards, and AI summaries, but no real moat.

RSAC 2026 made one thing very clear to me: the market is moving fast, but it is still deeply confused. The big announcements from Google, Splunk, and Databricks all point in the same direction. Security operations are becoming more agentic, more API-driven, and more automated. But most of the category still looks crowded, early, and only lightly differentiated.

The interesting part is not that everybody now has an AI story. It is where the pressure is landing: attack speed, active response, and the possibility that AI itself becomes the primary user of the security stack.

TL;DR

• Attacks are now fast enough that human-speed response is no longer a sufficient default.
• That will push the market toward active response, which is useful but also dangerous if the control logic is not deterministic enough.
• Most AI SOC vendors still sound similar because many of them sit on top of existing SIEMs and alert streams rather than changing the underlying detection or data architecture.
• The big SIEM vendors are moving, and one major EDR/SIEM vendor is expanding AI security into on-prem and sovereign environments.
• If AI becomes the user of security products, the UI matters less, the API matters more, and the economics of expensive SIEM platforms get harder to defend.

Attacks are getting faster

This is the part of the market I think people are still underestimating. CrowdStrike’s 2026 threat report says the average eCrime breakout time dropped to 29 minutes in 2025, and the fastest case it observed was 27 seconds. Databricks used its Lakewatch announcement to make a related point from the vulnerability side, citing research that mean time to exploit has fallen from 23.2 days in 2025 to 1.6 days in 2026.

That changes what matters in the SOC. A lot of SIEM workflows still assume there is time to search, enrich, discuss, and decide. That model was already strained. It gets worse when attacks speed up and when the adversary is using AI to compress its own loop. Search still matters, but a search-centric operating model is not enough if the environment can be compromised end to end in under an hour.

The obvious answer is more active response. The problem is that this is where things get dangerous. If teams start handing more containment and remediation decisions to AI before the systems are ready, we are going to see more self-inflicted outages. The market is moving there anyway, because the alternative is to keep defending at human speed against machine-speed attacks. SOAR was supposed to close part of that gap and clearly did not.

AI SOC is still confusing and mostly sounds the same

That was probably my main emotional reaction leaving RSAC: confusion. There were simply too many vendors with very similar messaging. RSAC says the conference had more than 600 exhibitors this year. I could not independently validate an exact count of 36 AI SOC vendors from public RSAC data, but “roughly three dozen” felt directionally right from the floor, and many of them sounded remarkably similar.

The common pitch was familiar: reduce alerts, triage faster, investigate faster, give the analyst a copilot, automate parts of response. Some of that is clearly useful. But a lot of it still feels like a layer on top of the existing SIEM rather than a rethink of the detection stack itself. If the AI mostly sits on top of alert streams coming out of a legacy backend, then it may improve analyst productivity without materially fixing false negatives, brittle detections, or poor data design upstream.

That is also why I do not think most of this market is really using LLMs in a deep way yet. In most cases, the models are being used for triage, recommendations, summarization, and analyst assistance. That is very different from using LLMs for real detection, broader SOC operations, or meaningful changes to the underlying architecture.

For a more complete framework of where AI SOC and SIEM should be heading, see raffy.ch/SIEM.

That is why so much of the category feels undifferentiated. The interfaces are different, the branding is different, and the demo flows are different, but the center of gravity often looks the same. The latest platform announcements only reinforce that point. If the platform owner adds the agentic layer too, the vendors sitting on top of Chronicle, Splunk, or similar platforms have a much harder moat to defend.

The architecture is shifting

By this point, the vendor movement is established. The more interesting question now is what it does to architecture. SentinelOne adds another signal here by pushing more AI security capability into on-prem, sovereign, and air-gapped environments.

Put together, that points to a broader market shift. Storage matters more. Data routing matters more. Sovereignty and local control matter more. Cheap data lakes, strong analytics layers, and flexible orchestration matter more. Traditional SIEM UI matters less than it used to, and that matters not just for SIEM vendors but also for MDRs that differentiated by putting an AI layer on top of someone else’s backend.

That is also why Splunk’s cost model keeps coming back into the conversation. Splunk is powerful and mature, but if the agent becomes the main consumer of the system, customers start asking a different question: am I paying for the analytics engine, or am I paying for UI, workflow, and operating complexity that an agent increasingly does not care about?

If AI becomes the user, the stack changes

The most important implication may be economic, not just operational. Security products were built for human analysts. The value lived in the UI, the workflow, the search language, the dashboard, and the services needed to make all of that usable. But what happens if the real user becomes Claude Code, Codex, Gemini, or some internal agent instrumented across the entire security stack? Daniel Miessler has been arguing that companies and products increasingly become APIs. Security looks like one of the clearest versions of that shift.

In that world, every product starts to look more like an API than an application. That is exactly where the recent announcements are heading. LimaCharlie’s new lc-soc release is a concrete implementation of the same idea: an open-source “agentic SOC as code” where AI agents are coordinated through the cases system and D&R rules, then deployed and versioned like infrastructure.

If AI becomes the primary user, the UI does not disappear, but it stops being the center of gravity. The agent does not care about your console. It cares about whether the data is accessible, whether the schema is consistent, whether the analytics layer is fast, whether the permissions model is clean, and whether the actions are safe to orchestrate.

That creates real pressure on expensive SIEM economics. If the agent can query multiple tools directly, the premium attached to a deeply monetized UI gets harder to justify. The market may move toward something simpler: cheap storage, a strong analytics layer, and an orchestration layer on top. That does not mean incumbents disappear. It means their value proposition changes. If AI becomes the user, the winners may be the vendors with the best APIs, control points, and data access model.

Evals become part of the control layer

The next problem is trust and determinism. Once you push AI beyond triage and recommendations and let it make or recommend more consequential changes, you need a way to keep the system reliable. That is where eval loops come in.

I heard Josh Saxe make this point at RSAC in the context of AI-first infrastructure management: if agents are going to make changes in live systems, you need strong evaluation around them to keep behavior bounded and repeatable enough to trust. I think the same logic applies directly to security operations. The market is moving toward active response, but the models themselves were not built around strict determinism.

That means the answer is not blind autonomy. It is more likely a layered system where adaptive AI sits inside clearer control boundaries, with evals, policy, and deterministic automation around it. Evals stop being an AI engineering detail and become part of the security control layer itself.

I recently joined Tim Peacock and Anton Chuvakin on the Google Cloud Security Podcast to talk about SIEM, AI SOC, pricing, federated architecture, detection engineering, and why network telemetry is quietly becoming important again.

The short version is simple: SIEM is not dead. Calling it obsolete makes for good marketing, but it is not a serious thesis.

The new wave of AI SOC, SIEM, and pipeline vendors is not proving SIEM is dead. It is proving SIEM vendors left too many gaps open for too long.

The recent wave of AI SOC startups, pipeline vendors, and new SIEM entrants is a response to real pain in the market. They are not replacing SIEM. They are capitalizing on the gaps incumbent vendors left open.

TL;DR

• SIEM is not dead. Vendors just left too many gaps open.
• AI SOC often exposes those gaps more than it replaces SIEM.
• Alert reduction alone will hide false negatives.
• The real fixes are better routing, detection, context, and workflows.
• Network telemetry still matters more than the market narrative suggests.

The market is not replacing SIEM. It is rebuilding missing pieces.

They say they will reduce alert volume, improve detections, make investigations faster, lower storage costs, and simplify operations. None of that is new. Those were always core parts of the SIEM vision.

That is why so many of these new entrants exist. They found real gaps:

• Pricing that became too hard to justify
• Architectures that did not scale as well as they should
• Detection stacks that still require too much manual work
• Default content that creates too much noise
• Workflows that remain painful for analysts and service providers

This is why I do not buy the “SIEM is over” narrative. If incumbents fix these gaps, many point solutions lose their edge quickly.

AI SOC is mostly a patch on downstream pain

The strongest short-term value in the AI SOC market is obvious: too many teams, especially MSSPs and down-market security providers, are drowning in alerts. A lot of environments are running with default content, light tuning, and limited budget for customization. Large enterprises can afford deep implementation and constant refinement. Many managed providers cannot.

If a product makes the SOC quieter without improving coverage, you may not have solved the problem. You may have just converted visible false positives into invisible false negatives.

If a startup is solving alert overload by learning that the same service-account misconfiguration fires every morning at 8am and can safely be deprioritized, that is useful. But it is still a patch on bad upstream logic, and it often hides a second problem: false negatives. Once teams see fewer alerts, they assume the system got smarter. Sometimes it did. Sometimes it just got quieter. The real fix belongs closer to the detection layer, the correlation logic, the content, and the configuration model.

That is why I think a lot of the current AI SOC wave is temporary in its present form. Not temporary because the need goes away, but temporary because the best parts of that value will be absorbed elsewhere. Some of it should move back into the SIEM. Some of it should live in the detection engine. Some of it belongs in better onboarding, better rule tuning, better data handling, and better defaults.

There is still room for new winners here. But “we reduce alerts by 80%” is not a durable thesis by itself.

The architecture debate is not centralized versus federated. It is about access patterns.

In theory, pushing compute to where the data sits is attractive. In practice, the answer depends on access patterns.

Some data absolutely does not need to be centralized all the time. Endpoint system calls are a good example. You do not want to shovel every low-level signal into a central platform by default if you can process, summarize, or prioritize it earlier.

But the moment an analyst, agent, or investigation workflow needs context, enrichment, and cross-correlation, some centralization comes back. You need to connect what happened on the endpoint with what happened on the firewall, identity plane, SaaS layer, email stack, and elsewhere.

So the future is probably not pure centralized or pure federated. It is hybrid:

• Keep some data local or near-source
• Route and centralize the parts that matter
• Pull deeper context only when needed
• Optimize around how investigations actually happen

This is why I keep coming back to smart data routing. Most organizations do not need to send every piece of data to the same place forever. But they do need an architecture that knows when to summarize, when to correlate, and when to pull more detail back in.

Data pipelines became the Trojan horse

Vendors in this space positioned themselves as optimization and routing tools. Send your data here, normalize it, trim low-value volume, route it to the right storage tier, keep costs down, and retain optionality. In many environments, that solved a real problem.

But the strategic consequence is bigger than cost control.

Once a pipeline vendor owns your ingestion layer and your integrations, it becomes an abstraction layer between you and the SIEM. That makes the SIEM less sticky. At first the pipeline vendor only routes data. Then it adds search. Then it runs lightweight detections. Then it supports simple rules. At some point it starts to look suspiciously like a simple SIEM.

If someone else owns the data path, they eventually get a shot at owning more of the security brain.

Pricing remains one of the category’s hardest unsolved problems

Almost everyone agrees that SIEM pricing has been a problem. Much fewer people agree on what the right answer is.

The vendor reality is straightforward: data volume drives cost. The customer reality is equally straightforward: they hate unpredictability.

That tension gets even worse in the service-provider world. MSSPs and MSPs often sell packaged services, per-user offerings, or per-device contracts. Their customers do not want a fluctuating bill because log volume spiked this month. So the thing that is economically clean for the vendor can be operationally ugly for the buyer.

There is no perfect answer here. But the next generation of pricing models will need to do a better job of separating:

• Predictable commercial packaging
• Actual backend resource consumption
• Incentives for better data quality rather than more raw ingestion

The market has already started experimenting. Bring-your-own-storage, bring-your-own-compute, lower-cost data lakes, and more selective routing are all responses to the same pressure. Pricing is one of the core forces reshaping the market.

Detection engineering still needs much more help from the platform

Rules still need adaptation by environment. Thresholds differ. Data quality differs. Sources differ. Customer expectations differ. Generic content does not simply drop in and work.

What is surprising is how much low-hanging product work still remains. A modern platform should do far more to help users answer basic but critical questions:

• Is the data required for this detection even present?
• Is it configured in a way that can ever make this rule fire?
• Are there obvious gaps or mistakes in the source configuration?
• Which detections are silent because they are poorly mapped to the environment?

The more interesting direction, in my view, is not just better standalone rules. It is better context. Call it a context graph, an entity graph, a risk graph, or something else. The naming matters less than the function.

You want a living model of users, devices, applications, identities, behaviors, and risk signals. If the system knows that a user is coming from their normal IP, on a familiar device, through a known browser pattern, after strong authentication, that should shape how other events are interpreted. If all of those signals change at once, that should shape the response differently.

That kind of context is where detection quality meaningfully improves.

Network telemetry is not “back,” but it is still critical

I do not think this automatically means a major standalone NDR renaissance. But I do think many teams went too far in treating network telemetry as secondary once endpoint and application visibility improved.

An endpoint is still a single point of failure. If you lose visibility there, the network can still tell you a lot. It can help validate what else is happening. It can show you unmanaged systems, OT environments, choke points, and traffic patterns you will not otherwise see clearly.

This matters even more now because some organizations are reassessing where systems and data live. In parts of Europe, I am seeing more discussion around data sovereignty, political trust, private clouds, and selective moves back toward local or regional infrastructure. As architectures spread and governance constraints tighten, network visibility becomes more important again.

So no, I would not frame this as “throw away EDR and buy NDR.” That is the wrong lesson.

What happens next

The real question is not whether SIEM survives. It is which vendors understand they are now selling data architecture, detection quality, analyst workflow, and decision support.

The SIEM market is heading into another rebuild cycle. Some AI SOC and pipeline startups will disappear, some will be absorbed, and some incumbents will finally fix what they should have fixed years ago. But the core need is not going away: security teams still need a place where signals come together, context gets built, detections improve, and response decisions get made.

That is still SIEM territory, even if the implementation looks very different from what we used to buy.

? If you are building, buying, operating, or replacing SIEM, I’d love your input. I’m collecting market data at raffy.ch/SIEM. Anyone can contribute, and everyone is welcome.

Update (2026-02): I released the SIEM Maturity Framework Workbook (v1.0) that turns this post into a practical scoring tool.

I have been talking to a few AI SOC and new SIEM market entrants over the past few weeks. I have voiced some opinions in previous posts but have now started to capture a list of features that I believe represent the openings existing SIEM players have created in the market for these new vendors to emerge.

Before I outline what I think those features are, let me be clear: this is my list. I am aware that existing SIEM vendors will claim that they already do many of these things. All I will say is this: market churn and capital flow suggest that these capabilities are either not as mature or not as integrated as claimed.

And to the AI SOC companies and investors: be careful about the short-term problems your investments are solving. Yes, there is real traction with MSSPs that are overloaded with false positives. And yes, many will gladly pay to reduce alert workload by 80%. But in many cases, these problems are being addressed superficially. Make sure you audit the underlying approaches and verify that the foundational infrastructure is sound. Solving this problem on top of an existing detection infrastructure doesn’t solve the problem at the core, which is the detections themselves. We need to fix those with some of the suggestions below to not needing a top-layer, alert reducer.

Without further ado, here are the items I am tracking. I welcome other opinions and additions to the list (no guarantee I will include them). Over the coming weeks, I will also try to rate some of the players across these categories to enable comparison. I could use help with that. Ping me.

A. DATA & CONTROL PLANE ARCHITECTURE

• Federation – The ability to query and reason over data where it lives, without forced centralization.
  (Another post following here at some point about the limitations of federation).
• Data Pipeline Optimization – Dynamic ingestion pipelines that enrich, route, sample, and filter data based on use case, risk, and downstream value. Not static “send everything to the lake.”
• Data Awareness – Understanding what data exists, what is missing, and what has silently degraded. The system must continuously reason about its own observability.
• Performance as a First-Class Constraint – Fast joins and low-latency queries across all relevant data. Real-time rule execution at scale. This is not about basic scalability, but about maintaining predictable performance as rule count and complexity increase, without simply throwing more compute at the problem.
• Modern AI Integration – The ability to integrate with emerging architectural patterns and frameworks, including MCP servers, vector stores, and related systems.

B. DETECTION & LEARNING SYSTEMS

• Hypothesis-Driven Hunting – Hunting should start with explicit hypotheses, not ad-hoc queries. These hypotheses should evolve, fork, and self-update based on outcomes. Agents swarms anyone?
• Automated Detection Tuning (Closed Loop) – Detections must evaluate their precision and recall over time. False positives and false negatives are signals. Humans stay in the loop, but are not the tuning engine. This also helps separate the detection engineering from the tuning that should be done by analysts.
• Environment-Adaptive Detections – Rules and models must adapt automatically to the specific environment, business processes, and user behavior and analyst feedback. Generic detections are table stakes.
• Detection Lineage and Memory – The system must remember why a detection exists, how it has changed, and what outcomes it has historically produced.

C. ENTITY-CENTRIC RISK & CONTEXT

• Asset Awareness – Effective protection and detection start with understanding what is being protected. Entity visibility is foundational: who owns this entity, what does it do, and which business processes does it support?
• Real-Time Entity Risk Scoring – Each entity has a continuously updated risk score driven by behavior, exposure, and contextual signals.
• Entity Risk Context – Risk is not a number. It is a set of properties that help explain the risk and provide context for decision making.
• Business Context Integration – Entities must be tied to business processes, ownership, and criticality, and this context must inform alert generation and prioritization. Some people have started calling this the Context Graph.

D. OPERATIONAL REALITY (SOC, MSSP, ENFORCEMENT)

• Simple Query Interface: Support for both natural language and structured query languages (such as KQL). Analysts need both.
• Alert Triage Automation – Using ‘advanced’ context to tune detections. Ideally we have business context available to continuously improve our detections.
• Blindspot Detection – The system must actively identify where detections cannot exist due to missing or degraded logs or logging configurations. This includes making sure that log sources are actually staying up and keep reporting what they have to.
• Real-Time Readiness for Enforcement – We need our systems to become preventative. Therefore, its risk model must operate in near real time. Attackers are acting too fast.

A Few Additional Comments for Context

This is not meant to be a SIEM RFP. I am intentionally not listing table-stakes capabilities such as basic scalability, data source support, or baseline detection depth.

This list is less about features than about where intelligence and control actually live in the system. I am also not being prescriptive on how these features are built. Many of them can benefit from AI / LLM / ML approaches and, in fact, should be using them.

Look at the list, then look at your AI SOC platform of choice. How much of the above does it truly cover?

If you are evaluating an AI SOC platform and most of its value proposition lives above alerts rather than below them, you should be skeptical.

Update (2026-02): I released the SIEM Maturity Framework Workbook (v1.0) that turns this post into a practical scoring tool.

Security has always moved in waves. Not because we suddenly get smarter, but because we learn from past mistakes, identify gaps, hit limits, need to protect new technologies, and then go and do our best to solve those new security challenges with the technologies at hand.

The era of AI (let’s be clear, we have had AI for a long time; what I mean specifically is the advent of Large Language Models) has shifted many industries, but specifically security in a particularly revealing way. AI did not just give us new tools to solve security problems. It invited innovators and entrepreneurs to revisit pretty much every security technology to see if LLMs could be useful to address some of the existing challenges. But that’s not where things stopped. More interestingly, some teams used this moment to question whether the underlying approaches themselves still made sense at all. Not just whether LLMs could help, but whether modern data architectures, different telemetry choices, and different enforcement models could fundamentally change outcomes.

That is what has triggered a real wave of new companies in cyber, including across markets that many considered mature, or even stagnant, like SIEM.

The Five Phases We Just Lived Through

Let’s take a non-scientific look at how major security approaches evolved over the past 25 years. This is not exhaustive, but it helps explain where we are today.

1. Network-Centric Prevention

Back, many moons ago, we started with firewalls, IDS, and later IPS. The model was simple. Look at packets. Stop bad things. It worked until attackers learned to look normal.

2. More Data, Centralized, Higher-Level Insights

When network telemetry created too many false positives, we added vulnerability data and authentication events and fed them into a SIEM to correlate. The results were “mixed”. Fortunately for the SIEM market, compliance and audit requirements emerged, mandating long-term log retention. This gave SIEM a durable justification, even when its security value was debated. SIEM became indispensable for visibility and forensics, but increasingly disconnected from real-time decision making.

3. Back to Prevention and Response

As SIEM alert volumes exploded and analysts could not keep up, the industry pivoted. EDR. NDR. SOAR. We all know how that played out. NDR never truly broke out. EDR became a major category. SOAR largely collapsed back into SIEM. And eventually, most large EDR vendors added a SIEM to their portfolio.

This was not convergence by design. It was convergence driven by operational gravity.

4. AI Triggers a Reality Check

LLMs made many believe they could simply layer AI on top of broken architectures. Some startups did exactly that. They will likely not be the long-term winners.

The more interesting group of companies used AI as a forcing function to re-examine first principles. What data actually matters? What can realistically be prevented at the edge? What must still be correlated centrally? What is structurally broken in SOC workflows? Where have we been compensating for bad architecture with human labor? Crucially, many of these answers have little to do with LLMs themselves, and much more to do with data fidelity, placement of control, and modern system design.
This is where the real innovation is happening.

5. The Convergence

We are now in a phase where prevention is moving back to the edge, while analytics and orchestration remain central. Endpoints are smarter. Browsers are instrumented. Networks are being re-observed. Context is finally treated as a first-class input.

But there is still a SOC. There is still a central nervous system that correlates, reconstructs, explains, orchestrates, and proves what happened. Call it SIEM, security analytics, XDR, or AI SOC. The name is irrelevant. The function is not.

In parallel, we are realizing that we can push enforcement / prevention back to the edge. Wherever we have enough information, execute at the edge. Where we don’t, call out to your central nervous system. To your brain. The brain (your SIEM) that understands at any moment in time, what the risk and function is of every entity in your network. And use that information for decision making.

Why AI SOC Will Collapse Back Into SIEM

Many startups brand themselves as “AI SOC”. What do they actually do?

They primarily ingest alerts from EDR, NDR, SIEMs, and cloud platforms, then attempt to determine which ones matter. They add context, apply behavioral analysis, and suppress false positives.

In other words, they attempt to do what SIEM, UEBA, and SOAR were always supposed to do, just with better math and more compute. However, there is one problem. Many of the AI SOC contenders operate on alert streams. That means they start from already lossy, opinionated data. Real behavioral analysis does not on top of alert streams. It lives in raw telemetry. Email flows. Network sessions. Browser actions. Endpoint system behavior.

Once an AI SOC platform decides to ingest that raw data directly, it immediately recreates the ingestion, normalization, storage, and correlation problems that SIEM already exists to solve. At that point, the separation no longer makes sense. This is exactly why UEBA and SOAR collapsed back into SIEM. And it is why AI SOC will do the same.

There will be one place where data is reconciled, correlated, and turned into decisions. That place will increasingly run on federated, near-real-time architectures rather than twenty-year-old indexing engines. But their function remains the same. Call it whatever you want. It needs to be one system, not many and it doesn’t care what you call it.

The Shift Is Not Just Technical. It Is Organizational.

What is interesting to note about these new entrants in the SIEM or security analytics space is not just their security architecture. It is the company architecture. Modern security startups are being built on AI-native operating systems: Sales calls are captured and analyzed, not just by sales, but product teams mine them for competitive signals, marketing uses them to refine messaging, engineering uses them to prioritize roadmaps. This is not a tooling upgrade. It is a fundamentally different operating model.

Imagine a system where the vision, mission, strategy, and priorities are centrally maintained, updated and codified. Every function consumes that shared intelligence to drive decisions, messaging, and execution. This does not just improve alignment. It dramatically compresses learning cycles and execution speed. And that, more than any individual feature, may be the hardest thing for incumbents to replicate.