Cisco is turning Webex into a workplace where AI agents can join the conversation, understand the shared context, and help move work forward. At WebexOne 2026 in Austin, Texas, on October 7, 2026, the company introduced new agentic collaboration products designed to connect employees, customers, AI agents, and business systems.
The goal reaches beyond adding another chatbot to a familiar workplace tool. Cisco wants teams to invite AI agents into spaces, meetings, and calls, then use them to execute complex, multi-step work across Webex and third-party applications. That means the conversation itself becomes the launch point for analysis, recommendations, scheduling, notes, translation, coaching, and coordinated follow-up.
Webex Becomes a Home for Collaborative Agents
Cisco said Claude Managed Agents will soon let teams move from shared team context to analysis, recommendations, and coordinated follow-up actions across approved agents without leaving Webex. Employees can invite collaborative AI agents directly into spaces, meetings, and calls, where those agents can handle work across Webex and third-party applications.
A personalized view will help employees prepare for upcoming meetings and calls while organizing their priorities. Agents participating in Webex spaces, calls, and meetings can help teams prepare for conversations, manage scheduling, take notes, translate discussions, and provide coaching.
That model also moves onto the phone. New Personal Agents for Webex Calling can answer when a user is busy, identify the caller’s purpose and urgency, and help with the next step. The agent does not simply respond inside a chat window; it becomes part of the interaction flow surrounding a call.
“The next era of work isn’t about bolting chatbots onto yesterday’s tools,” said Jeetu Patel, Cisco’s president and chief product officer. “It’s about AI agents working alongside people and teams, being deeply embedded in their workspaces, and collaborating on a secure, trusted foundation. When people and agents share the same context, work keeps moving and every customer relationship gets stronger.”
Cisco is also updating its Webex cloud collaboration platform to accommodate AI agents. A Webex Suite Model Context Protocol integration brings Webex context directly into OpenAI’s dots, and Cisco said dots will be available within Webex in the future as an always-on personal agent.
Snorre Kjesbu, Cisco’s senior vice president and general manager of collaboration, described the new AI-first Webex experience as a connected environment for people and agents. Cisco said it brings identity, security, governance, and visibility together so organizations can control what AI can access, what it can do, and how it acts.
Dialog Extends the Agent Push to Customers
Cisco’s new Dialog toolset takes the same agentic approach into customer interactions. The company calls Dialog an “agentic harness” that directs AI to handle customer interactions, assigning tasks to various agents that keep working on the customer’s behalf after a conversation ends.
Agents under Dialog’s umbrella can coordinate “across people, agents and backend systems.” That structure is designed to keep customer work moving after the initial exchange, rather than ending when a conversation closes. Dialog also speeds onboarding by treating “AI agents like new teammates.”
The platform “continuously refines agent performance with every interaction,” while Cisco promises a safety-forward experience through collaboration with Splunk, a data analysis company. Together, these pieces position Dialog as a system for coordinating customer-facing work across multiple participants and business tools.
The promise is big: AI agents that can carry context from a conversation into follow-up actions, work with people and backend systems, and keep pursuing an outcome for the customer. But the risks become bigger as workflows stretch across more steps.
The Reliability Test Behind the Hype
A study by Carnegie Mellon University found that even the best-performing AI agent in the test, Google’s Gemini 2.5 Pro, failed to complete real-world office tasks 70 percent of the time. The failure rate for multi-step tasks with AI agents sits between 60 and 70 percent.
Those numbers expose the challenge Cisco is trying to address with controls, shared context, and agent coordination. An agent may perform well on each individual action, yet a long workflow can still break when small errors compound across every step.
Even an agent with a 95 percent per-step reliability rate will succeed only around 36 percent of the time when handling a 20-step workflow. A 1 percent failure rate at each step produces approximately 36 percent success over 20 steps, making even a 99 percent per-step reliability rate risky when money and a company’s reputation are on the line.
That is why Cisco’s focus on identity, security, governance, and visibility matters as much as the ability to add more agents. Organizations need to control access and actions while seeing how agents operate across Webex, third-party applications, people, and backend systems.
Cisco’s announcement points toward a workplace where AI agents are not separate tools waiting for prompts. They are becoming participants in meetings, calls, customer conversations, and follow-up work. The next test will be whether these agents can turn that shared context into dependable results when the workflows become complex.
Based on
