Cisco unveiled a new generation of agentic collaboration capabilities on October 7, repositioning Webex around AI agents that work alongside people instead of waiting for prompts. The centerpiece is Claude in Webex, built with Claude Managed Agents, which lets teams analyze data, generate content, and execute multi-step work using shared enterprise context without leaving the collaboration app.
The announcement, made from Austin, Texas, addresses three connected dimensions of work for the AI era: how people work, how the workplace adapts, and how businesses engage with customers. The next era of work isn't about bolting chatbots onto yesterday's tools, Cisco president and chief product officer Jeetu Patel said. It is about AI agents working alongside people and teams, deeply embedded in their workspaces, collaborating on a secure, trusted foundation.
The practical vision is that digital teammates join the flow of work with the same context humans have. Intelligent workplace experiences unify secure networking, collaboration technology, and physical spaces, so an agent can correlate a meeting discussion with the state of the network or a customer record. That framing plays directly to Cisco's installed base: the company can sell an AI story that leans on the network it already manages for most large enterprises, rather than asking IT to buy a whole new agent platform from scratch.
The rollout also includes RoomOS 27 updates expected in November for Cisco devices, a Google Meet partnership so Cisco hardware can join rival meeting platforms more naturally, and a new customer-experience platform for contact centers. Coverage of the launch notes that the agentic push could strengthen Cisco's position in the enterprise collaboration market and open software and services revenue beyond its traditional networking hardware.
For enterprise buyers, the questions to ask are familiar ones from every AI launch: which agents get what data, how access is governed, and what happens to the shared context after the meeting ends. Managed agents that act across enterprise systems raise the stakes on identity, permissions, and audit trails, and Cisco's pitch is that the trust layer is built in. Evaluate the integrations, but scrutinize the governance model, because the difference between a helpful teammate and a liability is what the agent is allowed to touch.