Skip to content
Institute for AI TransformationInstitute for AI Transformation
Back to The BriefingAnnouncement analysis · Enterprise software

AI agents in Webex: who can authorize an action?

Cisco plans to bring Claude into Webex. Enterprise buyers need to understand what an agent can read, which actions it can take and when it must ask for approval.

The Institute for AI Transformation
Draft for review
Source announcement:
Copy article link
Editorial illustration of people and an AI agent sharing a collaboration workspace.
Institute illustration

What Cisco announced

At WebexOne on October 7, Cisco announced plans to bring Claude Managed Agents into Webex. It also described collaborative agents working across meetings, spaces and connected applications. The release identifies the Claude functionality as coming soon.

Source: Cisco · October 7, 2026 .

Institute interpretation

Questions for your evaluation

Cisco’s release says organizations will control what AI can access and do, but it does not explain the mechanism for granting, limiting or revoking authorization. A meeting request alone does not establish permission. Ask Cisco to demonstrate whose authority an agent uses and how a person approves, stops or reverses an action.

What to verify before deployment

Confirm availability for your account. Then test access controls, approval requests and the record of actions the agent takes.

Source and scope

This is an announcement of planned capabilities, not confirmation that every feature is available to every customer.

Agentic collaboration announcement . Reviewed October 17, 2026.