Security | Threat Detection | Cyberattacks | DevSecOps | Compliance

Compliance teams have control over approved corporate systems like enterprise software, managed databases, and internal applications. But they don’t have the same over what employees paste into ChatGPT, upload to Claude, or share with Gemini and other unauthorized AI tools. As such, when auditors review AI usage controls, most organizations discover they can’t prove that employees aren’t exposing regulated data through external AI services.

CrowdStrike is excited to announce Falcon OverWatch for Defender, a new offering that extends our elite managed threat hunting to Microsoft Defender environments. The need for proactive threat hunting is increasingly urgent as adversary operations evolve: 82% of intrusions observed in 2025 were malware-free, the CrowdStrike 2026 Global Threat Report revealed, and the fastest eCrime breakout time was a mere 27 seconds. Adversaries using AI increased their attacks 89% year-over-year.

Cybercriminals are getting smarter and faster. Social engineering attacks are evolving rapidly, and AI is making them more convincing than ever. According to the 2025 Verizon Data Breach Investigations Report, up to 68% of cyberattacks involve some form of social engineering. Meanwhile, 95% of cybersecurity professionals say AI is making phishing attacks harder to detect, and 65% believe attackers will soon rely on AI as their primary tool. This isn’t just theory.

Security operators often struggle with the escalating friction that naturally occurs in their detection and response (D&R) workflow. Detections fire in one tool. Investigations happen in another. Case tracking lives in a third. For MSSPs managing dozens of client environments, fragmentation compounds quickly. Analyst time bleeds into context-switching. SLAs are hard to track. When something goes wrong, reconstructing what happened across multiple platforms is painful.

If your first instinct when connecting an LLM to enterprise systems via MCP is to strip out all personally identifiable information, you’re building a system that is useless. The “block all PII” approach sounds responsible. It checks a compliance box. But it fundamentally misunderstands what MCP-based AI systems do and why they need data in the first place. The real engineering challenge is not blocking data.

Gartner named Zenity the company to beat in the AI Agent Governance category in its AI Vendor Race: Zenity Is the Company to Beat in AI Agent Governance report as of 17 April 2026. The evaluation covered technical capabilities, customer implementations, business model, and ecosystem strength. That methodology matters because for us, it means the recognition reflects what the platform actually does in production, not just how well a demo lands.

A recent open letter from the UK government on AI-driven cyber threats highlights a clear shift in the threat landscape. Cyberattacks are no longer constrained in the same way by human expertise, as advanced AI models can now help identify vulnerabilities, generate exploit code, and increase the speed and scale of attacks.

The Anthropic Glasswing initiative brings together Amazon Web Services, Apple, Broadcom, Cisco, CrowdStrike, Google, JPMorganChase, the Linux Foundation, Microsoft, NVIDIA, and Palo Alto Networks as launch partners. You can find a lot of posts and reactions on social media as it is definitely a big deal that Anthropic is keeping their Mythos Preview model out of general access.