Cisco Added Agentic Framework to Webex Platform
The update enables AI agents to coordinate across backend systems and persist on tasks beyond active user sessions.
Updated on Oct. 8, 2026 in Artificial Intelligence
Live Poll
Do you trust AI agents to successfully handle complex work tasks for you?
Cisco has integrated its new Dialog framework into the Webex platform, allowing AI agents to join meetings and manage customer interactions. These agents are designed to access internal knowledge bases and continue executing multi-step workflows even after a conversation concludes.
Why it matters
By enabling agents to bridge the gap between communication platforms and backend systems, Cisco aims to automate complex administrative tasks. This shift highlights an industry-wide focus on moving AI beyond simple chatbots toward persistent, autonomous office workers.
While researchers at Carnegie Mellon University report failure rates of 60 to 70 percent for AI agents on multi-step tasks, current systems often struggle with compounding errors. Even with a high 95 percent per-step reliability, completing a 20-step workflow results in only a 36 percent success rate.
The players
Cisco
A networking and communications conglomerate that develops the Webex enterprise collaboration platform.
Carnegie Mellon University
A research university known for its work in computer science and artificial intelligence.
Splunk
A software company focused on security and observability that provides tools for monitoring and securing enterprise data.
The details
The Dialog framework allows users to invite autonomous agents into Webex spaces to perform multi-step work by processing existing employee handbooks and operating procedures. These agents interface with backend systems to maintain continuity, meaning they continue working on assigned tasks after a call or meeting ends. Cisco has also integrated security features through a partnership with Splunk, a data platform for security and observability, to manage the data access required for these agents.
Timeline
Earlier in 2026: Carnegie Mellon University researchers conducted AI failure studies.
December 4, 2025: Social media posts highlighted AI agent failure rates.
The Tech Race
This integration marks a move by Cisco to operationalize autonomous agents within a standard enterprise stack. It follows a pattern set by the Carnegie Mellon University study, which quantified the significant performance hurdles facing AI agents tasked with complex office workflows.
Users will be able to invite AI agents into Webex calls and spaces to automate document-heavy tasks starting now. Enterprise adoption will rely on the agents' ability to effectively parse internal handbooks and navigate backend system permissions.
The takeaway
The effectiveness of the Dialog framework will likely be determined by its ability to maintain high success rates across long, multi-step sequences. Watch for future performance benchmarks released by enterprise software providers as they attempt to reduce the current industry-standard failure rates of 60 to 70 percent.
Further reading
For more on the current capabilities of autonomous systems, explore our Artificial Intelligence section.
Live Poll
Do you trust AI agents to successfully handle complex work tasks for you?
