What CrowdStrike's Own Press Release Actually Says
Published: September 18, 2026 | Category: Technology, AI | By Mahesh | Source: CrowdStrike's official press release, primary data current as of September 18, 2026
Today's new product is Falcon Guardian, which CrowdStrike introduced on September 1, 2026 at its own Fal.Con conference in Las Vegas. CrowdStrike's own press release opens with a framing that doubles as the thesis for the entire product: "AI hasn't changed the attack, it has changed its speed. Governance alone can't stop an agent already in motion. Falcon Guardian turns policy into protection, stopping threats where AI agents execute and before they can cause harm," according to CEO and founder George Kurtz.[1] Depth Grid has covered the proliferation of enterprise AI agents from the vendor side in Salesforce Gave Its AI Agents Job Titles Instead of Features; Guardian is the security industry's answer to the same trend, built for the IT and security teams now discovering they have no idea how many of those agents are actually running inside their own organisation.
The Problem This Is Actually Solving
Claim: A large volume of AI agents already operating inside real organisations are pulling instructions from sources nobody vetted, and some actively impersonate trusted AI labs. Source: Independent reporting on the broader wave of AI agent security launches at Fal.Con cites a new industry report that "found 17,800 public AI add-ons across 6.7 million installations drawing instructions from unverified external sources, including skills impersonating Anthropic and OpenAI that could run arbitrary code."[2] CrowdStrike's own press release frames the underlying risk directly: "as AI agents gain system-level privilege, the endpoint is where they reason, plan, and execute, accessing sensitive data and triggering downstream workflows with behavior indistinguishable from legitimate user activity."[1] CrowdStrike separately cites its own 2026 Global Threat Report showing an 89 percent increase in attacks by AI-enabled adversaries over the past year, alongside third-party CDW survey data that 43 percent of respondents had experienced phishing attacks enhanced or generated by AI.[3] Analysis: The most important word in CrowdStrike's own framing is "shadow," a term borrowed directly from the earlier "shadow IT" problem of employees adopting unsanctioned software, now applied to AI agents specifically. The scale here is the actual news: security teams are not dealing with a handful of unauthorised chatbot extensions, but tens of thousands of add-ons circulating with system-level access and no vetting process behind them, including ones deliberately designed to look like they come from trusted vendors. Published: September 1, 2026.
What Guardian Actually Does, Feature by Feature
Claim: Guardian's core function is connecting an AI agent's actions on a device directly to the security telemetry CrowdStrike already collects from that same endpoint. Source: CrowdStrike's own press release lists four capabilities available at launch. "AI Agent Discovery and Inventory" means "the Falcon sensor discovers known and shadow AI agents across Windows and macOS, providing a live inventory of every running and dormant agent across the enterprise, who deployed it, and its security status."[1] "Agent Runtime Visibility" "connects AI agent behavior directly to Falcon endpoint telemetry, establishing a causal chain from user prompt, identity, tool call, and skill use to every downstream system action, revealing the full agent execution graph."[1] "Agent Access Controls" "defines which AI agents are permitted to run on managed endpoints, blocking unauthorized agents and translating governance policy into enforceable runtime controls."[1] "Runtime Detection and Response" "detects attacks on agents and malicious agent behavior, reconstructs the full execution chain, and determines blast radius in real time."[1] CrowdStrike President Michael Sentonas summarised the shift in plain terms: "once AI becomes part of the enterprise, visibility is no longer enough. You need to know what's running, what it can access, what it is doing, where to draw the line. That is Guardian."[4] Analysis: The "causal chain" language is doing real technical work here, not just marketing. Most existing AI governance tools can tell an administrator that an agent exists or log an individual prompt in isolation, but connecting a specific prompt to the specific tool call, file access and system action it triggered, in sequence, is a genuinely harder correlation problem, and it is the piece that determines whether a security team can actually reconstruct what an agent did after something goes wrong rather than just knowing that something happened. Published: September 1, 2026.
The Endpoint Bet, and Why It's Structurally Self-Serving
Claim: CrowdStrike is explicitly betting that its existing dominance in endpoint security software gives it a durable structural advantage in AI agent security specifically, a claim tied directly to its installed base rather than to any AI-specific innovation. Source: The press release states plainly: "deployed across hundreds of millions of devices, CrowdStrike has more endpoint real estate than anyone. That structural advantage now defines AI agent security," and separately, "the endpoint is the only enforcement point with complete execution visibility, and where runtime security begins."[1] Kurtz draws the parallel to CrowdStrike's own founding thesis directly: "CrowdStrike pioneered EDR by making the endpoint the control point for stopping attacks. AI demands the same approach."[1] Analysis: This is a genuinely reasonable technical argument, since an AI agent running locally on a device ultimately has to touch the same operating system, files and processes any other software does, meaning existing endpoint telemetry is a real asset. But it is also, transparently, an argument that happens to favour the company making it: CrowdStrike is selling continuity with its existing product line and installed base, an approach a competitor without that same endpoint footprint would have every incentive to frame differently, for instance by arguing that agent security needs a purpose-built, network-level or model-level approach instead. Readers evaluating this claim should weigh it as a genuine architectural position with real merit, while recognising it is also the framing that most directly protects CrowdStrike's existing competitive moat. Published: September 1, 2026.
What's Announced but Not Actually Here Yet
Claim: Several of the capabilities in CrowdStrike's own launch announcement were not actually available at launch, and CrowdStrike disclosed no pricing or general availability date for the product's core functions. Source: CrowdStrike's own press release lists three additional features using future tense rather than present tense: "AI Gateway: Will provide a centralized control point for enterprise AI traffic across supported AI models and services"; "Falcon Complete for Guardian: Will deliver 24/7 expert-led detection, investigation, and response for AI agents"; and "Falcon Adversary OverWatch for Guardian: Extends managed cross-domain threat hunting... to AI agent activity."[1] Independent analysis of the launch confirms this gap directly: "three announced components were not available at launch and remain forthcoming," and separately notes "CrowdStrike did not disclose pricing or a general availability date for Falcon Guardian's endpoint-focused capabilities in the announcement."[4] Analysis: This is a common pattern in enterprise security launches, announcing a full platform vision at a flagship conference while shipping a subset of it, but it is worth stating plainly rather than letting the announcement's confident framing obscure it: roughly three of the seven capabilities described in CrowdStrike's own press release remain future commitments, not shipped features, and any organisation evaluating Guardian today should confirm current entitlement and actual feature availability directly with CrowdStrike rather than assuming the full press release describes what a new customer can deploy on day one. Published: September 1, 2026.
How to Actually Try It
CrowdStrike's own press release points prospective users to the Falcon Guardian product page and its companion blog post explaining the product's positioning in more depth.[1] CrowdStrike offers a free 15-day trial of the broader Falcon platform, and organisations wanting pricing details or a guided walkthrough can use CrowdStrike's own contact page, since, as noted above, Guardian's specific pricing was not disclosed in the launch materials.[1][4]
Common Questions
Sources
- CrowdStrike, "CrowdStrike Unveils Falcon Guardian to Secure AI Agents Where They Execute: On the Endpoint at Runtime," official press release, September 1, 2026. crowdstrike.com
- AI Agent Store, "AI Agents News: Week of September 13, 2026," citing a shadow AI industry report, September 2026. aiagentstore.ai
- BizTech Magazine, "CrowdStrike Fal.Con 2026: CrowdStrike Announces Falcon Guardian and Safe Mind for AIDR," citing CrowdStrike's 2026 Global Threat Report and CDW survey data, September 2026. biztechmagazine.com
- NAND Research, "CrowdStrike Falcon Guardian: AI Agent Security at the Endpoint," September 2026. nand-research.com
Read More on Depth Grid
- Salesforce Gave Its AI Agents Job Titles Instead of Features
- OpenAI Will Now Rent You the Same Agent Infrastructure That Runs Codex
- Anthropic Just Competed With Its Own Biggest Customers
- Enterprise AI Agents: Where the Big Checks Go
- Meta's New AI Agent Wants to Book Your Travel and Bills
Article by Mahesh | Depth Grid
Discussion (0)
Comments are disabled for this post.