Itdaily - KnowBe4 links Agent Risk Manager to Anthropic Claude Compliance API

KnowBe4 links Agent Risk Manager to Anthropic Claude Compliance API

Magnifying glass over the KnowBe4 website showing the logo and the slogan 'Human error. Conquered.'.

KnowBe4 is integrating Agent Risk Manager with Anthropic’s Claude Compliance API to monitor and secure AI agents in real time. The link is intended to improve governance now that AI agents are increasingly working autonomously with corporate data.

KnowBe4 announces that Agent Risk Manager, its security product for AI agents within the KnowBe4 platform, will henceforth integrate directly with Anthropic’s new Claude Compliance API. This allows organizations to more centrally monitor, analyze, and manage AI agent activities via a single operational dashboard.

The announcement fits into a broader shift where autonomous AI agents are increasingly performing multi-step workflows, gaining access to shared storage, and interacting with essential business data. According to KnowBe4, this creates security blind spots, especially when AI usage falls outside official oversight.

Why this integration is arriving now

According to KnowBe4, a growing number of organizations are deploying AI agents without security teams always having visibility into exactly what those agents are doing. In the ‘From Agentic Risk to Human Wins 2026’ report, 58 percent of cybersecurity leaders say AI agents are actively performing tasks within their organization, while 52 percent indicate that AI usage is not officially approved or supervised. This tension between adoption and control makes governance around AI agents a concrete operational problem.

KnowBe4 positions Agent Risk Manager as a way to restore that visibility. By configuring a single compliance API key, the product can centralize Claude deployments into one dashboard and assign activities to a responsible employee. The same overview should also help assess risks alongside other AI ecosystems such as Microsoft Copilot Studio, ChatGPT, Google Gemini, and so-called shadow AI solutions.

What Agent Risk Manager detects and logs with Claude

The integration focuses on context-aware threat detection. Agent Risk Manager’s detection agents assess Claude activities for prompt injection, exposure of sensitive and personal data, privilege escalation, excessive autonomy, and unauthorized access, among other things. KnowBe4 emphasizes that the analysis uses the full context of an interaction rather than just matching keywords, in order to better identify signals of abuse or policy violations.