AI agents are beginning to make changes to production networks, raising a practical question for the partners who manage them: how much autonomy can customers safely allow?
Cisco has released a report titled “The Impact of Agentic AI on Network Operations” which says that AI is moving from helping IT teams diagnose problems to taking operational actions itself. The study, independently conducted by Omdia Research among 1,000 IT and network-operations leaders at organizations with 500+ employees, found that 51% already have agentic AI acting in production network operations (NetOps), while 84% expect an AI-led operating model within 12 months.
Jim Frey, Chief Analyst, Network and IT Operations, at Omdia Research had this to say: “The findings point to a broader shift from the advisory model of AIOps to agent-powered operations, AgenticOps, that can take action.”
For MSPs and integrators, that shift could mean helping customers set limits on agent actions and check whether automated changes worked.
What the report says
Network operations teams face rising alert volumes and problems that span multiple systems. The report estimates that the average organization receives about 4,100 monitoring alerts and events a day, just over half of them network-related.
Beside the alert volume generated by traditional infrastructure, AI processes are also now adding to the network load. In separate testing, Cisco found that a task performed by an AI agent generated up to 450% more total network traffic than the equivalent non-agentic task.
Even though there are a number of factors to drive organizations to use agentic NetOps, the report findings says that organizations aren’t ready to make the shift without clear governance mechanisms. Thirty-six percent of respondents said full observability, including action traces, summarized reasoning and post-action audits, is the minimum acceptable standard for agentic operations.
What this means for enterprises
Enterprises will need to change their IT operations and governance policies, if they are to work in the changing NetOps landscape.
The survey found that 95% of respondents saw at least one significant shortcoming in their existing non-agentic AIOps tools, while 86% said a single integrated platform was the most effective path forward. For MSPs and integrators, that makes visibility across customers’ existing tools a useful starting point.
Another important point to note is that the IT workforce will simply change rather than shrinking. Humans can establish policies and escalation pipelines, while agents focus on the high-volume operational tasks.
As network administrators and AI agents increase their collaboration on NetOps, enterprises should build visibility and auditing trails to ensure that the autonomous actors still meet organizational standards. The Cisco and Omdia Research report has already shown us that change is imminent; enterprises can begin to plan how they’ll age their NetOps processes. managetheir NetOps.
Read more: Cisco’s broader AI and partner strategy offers context on where its infrastructure and services priorities are heading.