Cisco Is Seating Claude in Webex Meetings as a Team Member
At WebexOne, Cisco said Claude, running on Claude Managed Agents, will join spaces, meetings and calls. The more interesting part is the governance layer Cisco wants to own.
The meeting bot is getting a promotion. At WebexOne 2026 in Austin this week, Cisco said Claude, powered by Claude Managed Agents, will be able to join Webex spaces, meetings and calls "as a team member": pick up the context of a conversation, work alongside the people in it, and help finish the work that comes out of it.
Cisco's own phrasing is that the feature is coming "soon." It has not given a date, and several reports covering the event place most of what was announced in early 2027. That caveat matters. The direction is clearer than the timeline.
From summarizer to participant
The first generation of meeting AI listened and summarized. Cisco is pitching something different: agents that act. In one example from the announcements, a team member @mentions an analytics agent in a Webex space to pull usage statistics, then asks a presentation agent to update a PowerPoint deck with those numbers. Agents in spaces, calls and meetings are also described as helping teams prepare, manage scheduling, take notes, translate and coach.
Claude is the headline partner, but Cisco is opening the platform broadly. Teams will be able to invite collaborative AI agents into spaces, meetings and calls to run multi-step work across Webex and third-party applications. A Webex Suite MCP integration with OpenAI's dots lets users bring Webex context into dots, which Cisco says will become a Webex "digital teammate." Agents will also be reachable during voice and video calls through an AI Assistant side panel.
"We aren't building another chatbot," Cisco wrote in its announcement. "We're moving from using AI to working with AI."
The stack Cisco is building around it
Several pieces sit under the agent pitch:
- AI Agent Studio, for building, deploying and managing voice and digital agents across Cisco's collaboration portfolio.
- AI Agent 360, described as a single control plane for building, testing, monitoring and optimizing agents, with security and observability.
- Personal Agents for Webex Calling, which screen calls, assess urgency, prioritize and act for the user.
- Webex Translator Agent, already in controlled availability, offering real-time speech translation across 10 languages in the Webex app and Cisco 9800 Series desk phones.
Cisco's reach is the reason this matters. It says it powers communications for more than 82 million users across environments, including more than 20 million on Webex Calling and about 30 million on on-premises Unified Communications Manager.
Pricing: the agents are an add-on
Some AI features remain included in Webex Suite and Enterprise plans. Collaborative agents, Translator, Polling and Calling agents will require a new add-on, according to coverage of the event. Webex Suite starts at $25 per user per month; the add-on price was not published.
That structure is the industry pattern. Basic AI becomes table stakes in the seat price; agents that do work become a second line item.
The real product is governance
The most consequential part of the announcement may be the least flashy. An agent that joins a meeting can hear things, see shared screens and, if connected to other apps, change things. Someone has to decide which agents get into which rooms, what data they can see, and who approved that.
Cisco's answer is to make the collaboration platform the control point. "You set the limits. You define what agents can see," the company wrote. It is pairing that with Webex Zero Trust coverage extended to meetings, messages, recordings and whiteboards; deepfake detection with Pindrop for synthetic voice and video in meetings and calls; a compliance hub with Theta Lake; and post-quantum-safe cryptography.
Analysts at the event raised the practical questions administrators will face: who approves an agent's access to a space, and can that approval be scoped to a single team rather than the whole tenant? Those questions decide whether agents in meetings are a productivity gain or a data-exposure incident waiting for a calendar invite.
Why Anthropic wants the seat
For Anthropic, Webex is distribution into a workflow its models do not otherwise touch: live conversation in large enterprises. Claude Managed Agents gives Cisco a hosted way to run Claude with tools and state, without Cisco operating the agent runtime itself. For Cisco, putting a frontier model in the room keeps Webex relevant against suites that bundle their own assistants.
The read
The interesting shift is not that a model can join a call. It is that the meeting itself becomes a workspace with members who are not people, and enterprise collaboration vendors are racing to be the place where those members get admitted, scoped and audited. Cisco is betting that whoever runs the room gets to set the rules for every agent in it, including the ones it did not build. When this ships, the admin console will matter as much as the model.
