Cisco Turns Webex Into a Teammate, Not Just a Tool

On October 7, in Austin, Texas, Cisco opened its WebexOne 2026 conference with an argument about the future of office software. Meeting apps spent the last three years bolting AI chatbots onto the side of the chat window. Cisco now wants the agents to sit at the table: invited into spaces, meetings, and calls, where they pull data, build slides, and keep working after everyone has logged off.

The announcement centers on the new Webex, redesigned around what Cisco calls digital teammates. These are specialized AI agents that participate directly in the flow of work instead of waiting for a prompt. Employees summon them with a normal @mention, the same way they would flag a colleague, and the agents execute multi-step workflows across Webex and third-party applications.

The worked example Cisco gave is telling. A team member tags an analytics agent in a Webex space to pull usage statistics, then asks a presentation agent to update a PowerPoint deck with those numbers. The analytics agent knows the context of the conversation; the presentation agent knows the deck. Neither has to be copied anything by hand.

The agent that sits in the meeting, not beside it, is the direction the whole industry is moving.

Claude is coming inside Webex

Anthropic is a named partner. Cisco said Webex will support Claude Managed Agents, which let teams turn shared context into analysis, strategic recommendations, and coordinated action across approved agents. The feature is marked "soon," not shipping yet, which is worth noting: this is a roadmap announcement as much as a launch.

OpenAI is in the picture too. Cisco is bringing Webex context into OpenAI's "dots" through a Model Context Protocol integration, and in the future dots will appear inside Webex as an always-on personal agent. Cisco is the only collaboration vendor courting both Anthropic and OpenAI in the same announcement, which says something about where the leverage now sits. The model labs want enterprise distribution, and the meeting platform vendors want frontier models.

A digital "chief of staff" feature rounds out the software: an agent that surfaces high-priority items across the day. Cisco wrapped all of it in enterprise trust controls covering identity, security, governance, and visibility, so organizations decide what AI agents can access and what they are allowed to do.

The office itself gets sensors and blueprints

The second half of the announcement is about physical space. Cisco's Intelligent Workplace Experiences framework unifies secure networking, collaboration hardware, and physical spaces into a single architecture managed through Cisco Cloud Control. The pitch is that the office should sense, reason, and act, not just house desks.

The most concrete tool is Workplace Designer, a visual planning interface that can turn a street address into a tailored IT and facilities blueprint. Workplace Experience Metrics and Workplace Insights give IT teams room-level telemetry, correlating Wi-Fi coverage, network assurance, and real-world occupancy against room booking data. The practical point: companies can stop inferring office use from badge swipes and start measuring which rooms are actually used, by whom, and whether the network holds up when they are.

For corporate real estate teams sitting on expensive half-empty offices, that telemetry is the whole game. Every return-to-office policy since 2022 has suffered from the same information gap: managers know how many people badged in, not whether the collaboration spaces work. Cisco is selling the measurement layer that makes hybrid policies enforceable and defensible.

RoomOS 27 and the hardware bet

The hardware layer gets RoomOS 27, the next operating system for Cisco's collaboration devices, scheduled for release in November. It brings the Cisco AI Assistant onto the room hardware itself, with natural-language interaction in 32 languages. A spoken "the room is too warm" can lower the blinds and adjust the temperature, a small example with a large implication: the meeting room becomes an interface, not a box.

RoomOS 27 also introduces a Director Agent that manages camera framing and microphone balance dynamically, and adds Workspace Advisor and Smart Room Assurance to Cisco Cloud Control for investigating and improving spaces at scale.

The strategic headline, though, is interoperability. Cisco struck a partnership with Google to run fully featured Google Meet natively on Cisco devices, joining Webex, Microsoft Teams, and Zoom. Cisco is explicitly pitching its hardware as the most open collaboration hardware on the market: run whatever meeting platform you like, and the Cisco device layer manages it all the same way. For enterprises that run two or three meeting platforms at once (most do), that openness removes a purchasing argument for rivals.

Capping the hardware list is the Ceiling Speaker Pro, an IP-connected ceiling speaker that uses AI to adapt audio levels across the space. It connects with a single Ethernet cable and joins the same managed system as Cisco cameras and microphones, replacing multi-vendor audio setups with one network-managed stack.

Dialog and the long-horizon agent

The third pillar is Dialog, Cisco's agentic harness for the Webex AI Agent platform, aimed at customer relationships rather than internal meetings. The idea is to replace isolated, per-interaction customer service with continuous relationships: agents that keep working on a customer's behalf after the conversation ends, coordinating across people, agents, and backend systems until the outcome is achieved.

Dialog keeps persistent relationship context across channels and time, self-learns from each interaction within enterprise policy, and onboards by indexing existing handbooks and knowledge bases. The guardrails matter here: Webex AI Agents operate under Cisco AI Defense with live telemetry and observability powered by Splunk, which Cisco owns.

Long-horizon execution is the genuinely new capability. Most enterprise agents today die when the chat window closes; an agent that persists across systems and time is a different liability and a different value proposition. Whether companies will trust an agent to keep working unsupervised is the question Dialog has to answer in the field.

What this means for everyone else

Zoom and Microsoft are on the same trajectory with their own AI companions and Copilot agents, and the pattern is converging: every major meeting platform is becoming an agent host. Cisco's differentiation is the combination of three things, agents in meetings, sensors in the office, and openness to every meeting platform including its competitors.

As with any vendor conference, the timeline language matters. "Soon" and November releases mean the agents and the operating system are commitments, not deployments. The meaningful test will be whether the Claude integration and the @mention agents do multi-step work reliably in real enterprises, with real governance, at scale. If they do, the meeting app of the next decade looks less like a window and more like a colleague that never leaves.