Claude Cowork Just Expanded Your Attack Surface. Is Your Security Ready?

November 20, 2026

·

8:00 AM PDT

Claude Cowork Just Expanded Your Attack Surface. Is Your Security Ready?
Play video

Register to watch

Claude’s latest update combines chat and agentic work, enabling new capabilities by default for eligible enterprises unless explicitly disabled. Join PointGuard AI to learn what to review now, close gaps in permissions and oversight, and avoid recurring security fire drills. Discover how proactive visibility, enforceable controls, and governance help your enterprise own its AI trust boundary as providers keep changing the rules.

Address the immediate exposure and build reliable backstops for whatever changes next.

Claude’s latest update brings chat and agentic work into one experience, with default enablement for eligible Enterprise organizations that have not explicitly kept it disabled. For security teams, the immediate questions are practical: What changed in our environment? What can AI access and do? Which permissions, approvals, and monitoring need attention now?

This is also a broader warning about relying on a provider’s defaults to define your security posture. As AI providers combine chat, tools, browsers, and autonomous workflows, enterprises can inherit risks their original review never considered, turning yesterday’s security review into today’s blind spot. Join PointGuard AI to explore how to close immediate gaps and replace recurring fire drills with proactive AI security. Enterprises must own and set the boundaries of AI adoption, with reliable visibility, enforceable controls, and governance that keeps pace with changing tools and autonomy.

What you’ll learn

We’ll start with the immediate response, then show how to establish lasting control across your AI environment:

•  Identify what changed and prioritize securityand access reconfiguration

•  Discover AI usage, applications, agents, andthe tools and data they can reach

•  Set permissions and human approvals accordingto task, sensitivity, and trust

•  Monitor runtime behavior for actions thatexceed an agent’s intended scope

•  Interrupt unsafe actions with graduated controls, including circuit breakers and kill switches

•  Build reliable backstops for changing AI tools, permissions, and autonomy

Watch the full recording

Complete this form to unlock the on-demand webinar.

Thanks for registering. Click the video to start watching.