Summary
Anthropic added inference hooks in beta for Claude Enterprise, letting compliance teams enforce policy in real time. Each governed prompt across claude.ai, Cowork, and Claude Code is held for an organization's AI security server to return allow or deny before inference proceeds, with signed requests and every denial logged in the compliance Activity Feed.
What changed
On August 5, 2026 Anthropic introduced inference hooks (beta) for Enterprise organizations. Organizations point Claude at their AI security server, which returns an allow or deny verdict on each governed prompt before inference runs; requests are signed, failure handling is configurable, and denials are recorded in the compliance Activity Feed.
Why it matters
Enterprises have hesitated to broaden LLM access without inline control over what leaves their boundary. A pre-inference allow/deny hook gives compliance teams a single enforcement point spanning Claude's chat, coding, and agent surfaces, removing a key adoption barrier for regulated organizations.
Evidence excerpt
Inference hooks are now in beta for Claude Enterprise organizations. Point Claude at your organization's AI security server, and each governed prompt across claude.ai, Cowork, and Claude Code is held for the server's allow or deny verdict before inference proceeds.