Cisco's agentic Webex updates: what is available now and what is still planned

These examples describe intended functions; the reporting does not document a live test of either task or measure whether the agents improve employee productivity.

Cisco also announced Dialog, a software layer for its Webex AI Agent platform. Cisco says Dialog is designed to retain customer context across channels and allow agents to keep working after an interaction ends, coordinating with people, other agents and backend systems.

The company says organizations will be able to onboard Dialog using material they already have, including conversations, policies, handbooks and procedures. Cisco anticipates a beta in the first quarter of 2027; that is a forecast, not a release already under way.

The announcement reflects a broader plan for agents to work with people and with one another. Snorre Kjesbu, Cisco’s senior vice president and general manager of collaboration, told SiliconANGLE: “Connected intelligence goes from people to people but also people to AI and AI to AI.” That is Cisco’s description of its direction, not evidence that the announced tools can yet carry out those tasks reliably.

Which Webex features are available now?

Cisco says Webex AI Agent is generally available for its on-premises Contact Center Enterprise offering. That availability applies to the named product; it does not mean Dialog or the newly announced agents for team conversations and meetings are generally available.

Cisco planned to add voice support for Contact Center Express in October 2026, but the supplied reporting does not confirm that it has shipped. SiliconANGLE also reported that Cisco plans to make RoomOS 27, its meeting-room software, available in November.

For administrators, Cisco announced expanded management and observability for AI Agent 360 using Splunk technology. Observability tools provide information about system activity. Cisco says the planned expansion will connect agent behavior to operational data, including escalation patterns and business impact, with general availability planned for December 2026.

Cisco also described security capabilities through AI Defense. The reports confirm the announcement, but do not independently assess whether those capabilities prevent errors or misuse. The same limit applies to the announced oversight features: their planned functions do not establish how well they work in operation.

What do agentic and autonomous mean here?

In Cisco’s enterprise context, “agentic” refers to software intended to take a sequence of actions across connected tools or continue work after a customer interaction. “Autonomous” describes delegated actions within a business workflow; the reporting does not establish that the systems operate without human oversight in every situation.

That distinction matters because a task that crosses applications or continues after a call can involve more steps and systems than a single response. The supplied reports do not show how reliably Cisco’s agents will complete those steps, preserve customer context or follow an organization’s policies.

Vinod Muthukrishnan, the Cisco executive who leads Webex customer experience operations, told SiliconANGLE: “We are talking about securing, observing and governing an entire execution chain that starts with the conversation and goes well into the enterprise.” His comment describes Cisco’s stated oversight goal; it is not an independent evaluation of the announced controls.

What do the reports confirm, and what remains untested?

SiliconANGLE, Engadget, CRN and CX Today reported Cisco’s WebexOne announcements. Their reporting supports that Cisco announced the features and plans, but does not establish that the promised agents have been tested in operation. SiliconANGLE reported on the product plans and Muthukrishnan’s comments; Engadget also raised questions about whether multi-step agents will work as intended, without testing Cisco’s products.

The reports provide no independent results on the agents’ ability to complete multi-step tasks, preserve customer context or improve business outcomes. They also do not independently establish the effectiveness of Cisco’s security and monitoring features. Cisco’s examples remain product descriptions, not demonstrated performance results.

The next stated milestones are RoomOS 27 availability planned for November, expanded AI Agent 360 observability planned for December, and a Dialog beta Cisco anticipates in the first quarter of 2027. The reports do not specify what human-approval controls will apply when agents take consequential actions, or provide independent tests of how they perform in enterprise workflows.