Cisco AgenticOps revealed
- Cloud Managed Networks posted details about Cisco's AgenticOps, an AI-in-the-loop system for IT operations. - AgenticOps is designed to reduce alert noise, scale decision-making, and augment human judgment across networks and devices. - The announcement signals Cisco is productizing agentic workflows to help manage fleets of room devices and endpoints. (x.com)
Cisco is turning “agentic” artificial intelligence from a concept into a product line for network operations, with AgenticOps now positioned as an operating model across its portfolio. (newsroom.cisco.com) In Cisco’s description, AgenticOps uses artificial intelligence agents to watch telemetry, correlate signals, recommend actions, and in some cases execute tasks with oversight instead of stopping at alerts. Cisco said the model was first launched in 2025 and expanded on February 10, 2026 across networking, security, and observability. (investor.cisco.com) The basic problem is familiar to network teams: too many dashboards, too many alarms, and too much manual triage across cloud, campus, branch, and edge systems. Cisco says AgenticOps is meant to compress the gap between detecting a problem, understanding it, and fixing it. (blogs.cisco.com) Cisco ties that pitch to a data advantage it calls the Deep Network Model, which it says is trained on more than 40 years of Cisco operational data, including production telemetry and Cisco Certified Internetwork Expert knowledge. The company says that model is meant to give its agents more network-specific reasoning than a general-purpose large language model. (cisco.com) The new push also helps explain why Cisco has been folding management tools into a cloud-first platform. At Cisco Live in San Diego in June 2025, the company said the Meraki dashboard and Catalyst Center were being brought together to support unified management and AgenticOps across cloud, on-premises, and hybrid environments. (blogs.cisco.com) That matters for fleets of devices because the same operating model can be applied beyond switches and access points to endpoints that generate constant health, usage, and incident data. Cisco’s public materials describe AgenticOps as spanning campus, branch, cloud, and edge, which is the kind of footprint that includes room systems and other managed endpoints. (cisco.com) Cisco has started packaging that idea into named tools. In February, it highlighted Agentic Workflows and AI Canvas as interfaces for troubleshooting and automation, alongside wireless-specific AgenticOps features that it said can predict issues, recommend configurations, and handle routine work more autonomously. (newsroom.cisco.com, blogs.cisco.com) Outside coverage has framed the move as broader than a single feature launch. Network World reported on February 11 that Cisco’s expansion touches core networking, security, and data-center products, underscoring that the company is using AgenticOps as a cross-portfolio operating layer rather than a narrow assistant. (networkworld.com) Cisco’s own language is careful on autonomy. The company describes AgenticOps as “autonomous action with built-in oversight,” and its blog posts stress “AI-in-the-loop,” meaning humans still approve, guide, or constrain decisions instead of handing operations to software outright. (investor.cisco.com, blogs.cisco.com) The upshot is that Cisco is no longer talking about agentic operations as a future state. By April 2026, it has attached the label to shipping management, troubleshooting, and automation products meant to run large estates of networks and connected devices with fewer manual handoffs. (newsroom.cisco.com, blogs.cisco.com)
