Cisco is turning its work platforms into active collaborators. Its new agentic AI rollout spans Webex, workplace devices, RoomOS 27 and Dialog, a customer experience platform designed to keep helping after a conversation ends.
The shift reaches beyond chatbots that wait for a prompt. Cisco is building AI agents into meetings, calls, office spaces and customer journeys, while bringing identity, security, governance and visibility together so organisations can control what agents access, perform and how they behave.
Webex Moves From Assistants to AI Teammates
At the center of the rollout sits an updated Webex platform that moves from prompt-driven assistants toward AI teammates that take part in work. These agents can join Webex spaces, meetings and calls, then handle complex, multi-step tasks across Webex and approved third-party applications.
That changes what collaboration can mean. A team member could tag an analytics agent in a Webex space to retrieve usage data, then ask a presentation agent to update a slide deck with the results. The work flows through the same shared context instead of forcing people to move between separate tools.
New Claude Managed Agents will let teams turn shared context into deeper analysis, strategic recommendations and coordinated action across approved agents. Cisco is also integrating OpenAI dots through the Webex Suite Model Context Protocol, which feeds Webex context directly into dots. Cisco plans to make dots available inside Webex as an always-on personal agent in the future.
Jeetu Patel, Cisco president and chief product officer, framed the change as a new model for workplace technology: “The next era of work isn’t about bolting chatbots onto yesterday’s tools. It’s about AI agents working alongside people and teams, being deeply embedded in their workspaces, and collaborating on a secure, trusted foundation.”
That trusted foundation matters because agentic systems do more than generate answers. They can retrieve information, make recommendations and take action across connected services. Cisco says its approach brings identity, security, governance and visibility together, giving organisations control over what AI is permitted to access, perform and how it behaves.
Every Meeting Room Becomes Part of the AI Network
Cisco is also converging secure networking, collaboration devices, digital experience assurance and physical spaces through Intelligent Workplace Experiences. The architecture is managed through Cisco Cloud Control, creating one framework for the systems that connect people, rooms and office operations.
Workplace Designer now expands visual planning from individual rooms to entire sites, covering networking, collaboration and workplace intelligence. Experience Metrics connects network and device health with floor plans, Wi-Fi coverage and occupancy data, helping IT teams see how technical performance relates to the physical workplace.
Workplace Insights tracks office use over time and compares bookings with actual occupancy. That gives organisations a clearer view of how people use their spaces instead of relying only on reservations.
RoomOS 27, the latest operating system for Cisco Collaboration Devices, is due to become available in November. Through a new partnership with Google, it supports native integration with Google Meet. Cisco devices now run Webex, Microsoft Teams, Zoom and Google Meet natively, giving meeting spaces access to all four services.
The updated Cisco AI Assistant supports natural-language commands in 32 languages. Cisco Ceiling Speaker Pro adds another layer to the room experience as a fully IP-connected ceiling speaker that uses AI to adapt audio across meeting spaces.
Dialog Keeps Customer Relationships Moving
Cisco has also launched Dialog, an agentic platform built to move organisations from isolated interactions toward continuous, personalised relationships. Dialog’s agents keep working on a customer’s behalf after a conversation ends, coordinating across people, other agents and back-end systems.
The platform maintains persistent context from customer history and enterprise data. It refines performance by learning from customer journeys and human agent judgment, while allowing rapid onboarding. The goal is not just to resolve one request, but to carry useful knowledge into the next step of the relationship.
That ambition faces a hard reliability problem. Matthew Noe said: “The numbers remain dismal. Researchers at Carnegie Mellon University found earlier this year that even the best-performing AI agent, which was Google’s Gemini 2.5 Pro at the time, failed to complete real-world office tasks 70 percent of the time.”
The failure rate for AI agents in multi-step tasks sits between 60 and 70 percent. Even an agent with a 95 percent reliability rate at each step succeeds only around 36 percent of the time in a 20-step workflow, and a 99 percent per-step reliability rate still seems incredibly risky.
Cisco’s rollout places that challenge at the heart of its design. Webex agents, workplace intelligence and Dialog all depend on agents completing connected tasks without losing context or exceeding their authority. With RoomOS 27 due in November and AI reaching across collaboration, devices and customer experience, Cisco is betting that secure coordination can turn agentic AI from a promising assistant into a working part of the enterprise.
Based on
