Organizations are moving rapidly toward using AI to perform more autonomous functions in network operations, according to new research conducted by Omdia for Cisco.

The study,The Impact of Agentic AI on Network Operations, surveyed 1,000 IT and network operations leaders at organizations with 500 or more employees across North America, Western Europe, and Asia-Pacific. The findings examine AI adoption in network operations (NetOps), attitudes toward AI autonomy, and the conditions organizations consider necessary for trusting AI-driven operational decisions.

More than four in five respondents (84%) expect to reach an AI-led operating model within the next 12 months. Meanwhile, 75% report that they have already deployed AI for NetOps, and 51% say they currently have agentic AI operating in production.

The research describes a shift from AI systems primarily providing recommendations to AI agents that can take operational actions. Respondents reported using or considering AI for tasks such as rerouting traffic, adjusting wireless parameters, isolating suspicious endpoints, and resolving incidents.

Network Complexity Drives Interest in Automation

The findings come as network operations teams contend with increasing alert volumes, fragmented management tools, and issues that can span multiple technology domains.

According to the study, the average organization generates approximately 4,100 monitoring alerts and events each day, with more than half related to the network. The research estimates that manually addressing the daily network-alert backlog would require roughly 100 IT specialists.

The survey also found that 92% of respondents regularly encounter performance issues spanning multiple domains and requiring correlation across 10 or more tools. In addition, 57% said their existing change-management processes cannot keep pace with current operational requirements.

Ninety-five percent of respondents said their existing, non-agentic AIOps tools fall short in at least one significant area.

Autonomy Accompanied by Requirements for Oversight

The research indicates that organizations are increasingly open to allowing AI to perform operational tasks without direct human intervention. Eighty percent of respondents said they are comfortable granting AI a high or fully autonomous role in NetOps, including 24% who are comfortable with AI operating without human oversight.

Eighty-two percent said they are comfortable allowing AI to make at least some production network changes without prior human approval.

At the same time, the survey points to transparency and control as important conditions for expanding AI autonomy. Sixty-nine percent of respondents said detailed explainability for agent-driven actions is required, while 36% identified full observability—including detailed tracing, summarized rationale, and post-action audits—as the minimum acceptable standard.

Eighty-six percent said a single integrated platform represents the most effective path for managing these capabilities.

AI Workloads Add Another Layer of Network Complexity

The study also addresses the potential impact of AI workloads on network infrastructure. Cisco's analysis of aggregated direct-to-AI network telemetry indicates that AI-related traffic is on a trajectory to double every six months. Separately, Cisco testing found that tasks performed by agentic AI can generate up to 450% more total network traffic.

These figures are based on Cisco's own analysis and testing rather than the Omdia survey and therefore represent a separate set of findings associated with the report.

The research suggests that the move toward agentic operations will require organizations to address both the operational opportunities and the additional visibility, governance, and infrastructure requirements associated with increasingly autonomous systems.

"The striking takeaway is not simply that AI adoption in NetOps is growing, but how quickly organizations are preparing for AI-led operations," said Jim Frey, Chief Analyst, Network and IT Operations, from Omdia. "The findings point to a broader shift from the advisory model of AIOps to agent-powered operations, AgenticOps, that can take action. As that shift accelerates, integrated visibility, strong governance, and measurable outcomes will be critical to expanding autonomy while maintaining operational control."

"Our team isn't growing, but complexity keeps increasing with significant momentum," said Mark Rodrigue, Senior Network Engineer at Room & Board. "Deep reasoning in the Cisco AI Assistant is one example of AgenticOps for us. Questions that used to mean a manual hunt-and-click exercise across dashboards and multiple data sources now come back in minutes, with the executive summary first and the supporting evidence underneath. That's a force multiplier! It shows its work, so I can follow the logic and see the full evidence chain behind every recommendation. That's the trust it takes to deploy agents at scale."

For network operations teams, the research highlights a developing model in which AI agents may assume more responsibility for monitoring, troubleshooting, and network changes while organizations establish the visibility and controls needed to oversee those actions.

Source: Cisco Systems, Inc.

Stay Connected with ISE Magazine 

Subscribe to our newsletters and magazine for the latest telecom insights, explore the current issue for in-depth features and strategies, and register for upcoming webinars to learn directly from industry leaders.