Headline
80% of network professionals are comfortable giving AI a high or fully autonomous role in network operations, according to a new survey of 1,000 IT and network operations leaders.
Key Facts at a Glance
- 80% are comfortable granting AI a high or fully autonomous role in network operations.
- 56% will allow that only with human approval, while 24% are comfortable with AI taking network actions without human oversight.
- 82% are comfortable allowing AI to make some production network changes on its own for certain categories.
- 51% already use agentic AI tools in production to take corrective action in real time, not just for advice.
- 84% expect to reach a fully AI-led operating model within twelve months.
- 75% use AI in some fashion for network operations.
- 99% would not trust AI to act without strict guardrails, including explainable AI actions, human approval, policy-based operational limits, emergency override mechanisms, role-based access control, and immutable audit trails.
- 59% make changes to production network environments at least daily, and 57% say their change processes cannot keep up.
- 92% say performance issues cross multiple domains, including cloud, security, applications, and endpoints.
- 95% say existing non-agentic tools fall short, mainly due to too much human interpretation and lack of cross-domain visibility.
- Two-thirds say the generative AI boom has increased network complexity.
- Average daily AI traffic is on a trajectory to double every six months.
- The average organization generates about 4,100 monitoring alerts and events per day, with 51% network-related.
- One research estimate suggests a typical practitioner can review, investigate, and resolve about 21 network alerts a day, meaning it would take about 100 specialists to handle the daily volume.
- 46% of alerts are closed without investigation.
- 65% cite alert fatigue as a meaningful source of employee dissatisfaction, and 67% say alert volumes prevent teams from doing other critical work.
- Organizations rely on an average of 10 tools for end-to-end visibility, but these tend to be siloed.
- The mean time to resolve a network incident is 88 hours; the median is 12.5 hours.
Network Operations Are Hitting a Complexity Wall
Enterprise networks have become too complex for humans to manage alone. That is the central message from a new report based on a survey of 1,000 IT and network operations leaders. The research, conducted by a major networking vendor and a research and advisory firm, describes an environment where change is constant, problems cross multiple domains, and the tools meant to simplify operations often add another layer of fragmentation. Network professionals are now pinning their hopes on artificial intelligence to deliver relief from an uptick in complexity that is only getting worse, in no small part because of the increased use of AI itself.
The numbers paint a stark picture. More than half of respondents already use agentic AI tools in production to take corrective action in real time, rather than simply taking advice from them. Eighty-four percent expect to reach a fully AI-led operating model within twelve months. Three-quarters of those surveyed say they use AI in some fashion for network operations, and 80% are comfortable granting AI a high or fully autonomous role. More than half will do so only with a human approving those actions, but 24% are comfortable with AI taking network actions with no human oversight. The vast majority, 82%, are comfortable allowing AI to make some production network changes on its own for some categories.
The vendor behind the survey calls this move from AIOps to agent-powered operations, or AgenticOps. Longtime industry watchers are on board with the idea. One analyst noted that networking professionals will embrace agentic operations because they have little choice. Trust in the new tools may take time, similar to the curve for autonomous vehicles, which get into far fewer accidents than cars with human drivers. By the same token, agentic AI tools will make mistakes, but far fewer than people will, and they will let people focus on higher-value work.
Agentic AI Moves From Advice to Action
The shift from passive AIOps to active AI agents is significant. Traditional AIOps tools analyze telemetry and surface anomalies, but they still require humans to interpret the findings and decide what to do. Agentic AI goes further. It can diagnose a problem, choose a response, and execute it through predefined workflows. That is why 51% of survey respondents say they already use agentic AI tools in production for real-time corrective action. The appeal is obvious: with networks changing daily and alerts piling up, human teams cannot keep pace.
Yet autonomy raises the stakes. An AI agent that takes a wrong action on a production network can cause outages, security gaps, or compliance failures. That is why the survey found near-universal demand for guardrails. Ninety-nine percent of respondents said they would not trust AI to act without such controls. The required guardrails include explainable AI actions, human approval for actions, policy-based operational limits, emergency override mechanisms, role-based access control, and immutable audit trails.
Guardrails build trust by letting users see the reasoning that leads an agent to a conclusion. In addition, while AI agents can draw conclusions from the intelligence they examine, any actions they take map to predefined workflows based on the network team’s standard operating procedures for how to respond in different situations. Another key constraint is that agentic agents do not talk to other agents. That helps a lot because when they are constrained to only figuring things out themselves, they are a lot less dangerous. Agents are also built around specific skills or knowledge and a defined scope of responsibility. That can be controlled when the work is done on a product basis. When those elements are put together, they give enterprise network operations teams the confidence to leverage agentic AI to help manage increasing complexity.
Driving the Need: Complexity, Rapid Change, and AI Itself
Complexity is indeed increasing. Fifty-nine percent of survey respondents report making changes to their production network environments at least daily. Half of those organizations are making multiple changes per day, and for a meaningful share, change happens multiple times per hour. As a result, 57% say their change processes cannot keep up. The pace of change is not the only problem. Ninety-two percent of respondents say performance issues tend to cross multiple domains, including cloud, security, applications, and endpoints. Similarly, 95% say their existing, non-agentic tools fall short in significant ways, mostly by requiring too much human interpretation and lacking cross-domain visibility.
Two-thirds of respondents say the generative AI boom has also increased network complexity. Traffic analysis of direct-to-AI traffic shows average daily AI traffic is on a trajectory to double every six months. This acceleration is likely fueled by the growing complexity of AI tasks, which demand more data exchange than a simple query. More AI traffic means more bandwidth pressure, more latency sensitivity, and more security scrutiny. It also means more network changes, more telemetry, and more alerts.
Perhaps not surprisingly, the survey found the average organization generates around 4,100 monitoring alerts and events per day, with 51% being network-related. One research estimate suggests a typical practitioner can review, investigate, and resolve about 21 network alerts a day. That means it would take a team of about 100 specialists to handle the daily volume. Since few organizations have that much staff, nearly half of alerts, 46%, are closed without investigation. Alert fatigue is a meaningful source of employee dissatisfaction for 65% of respondents, while 67% say alert volumes prevent teams from doing other critical work. Independent research around security alerts has found similar patterns, with well under 50% of alerts investigated.
Too Many Tools, Too Much Time to Resolve Issues
Organizations rely on an average of 10 tools to try to maintain end-to-end visibility, but those tools tend to be siloed. That makes it difficult to diagnose problems that cross domains, as many do. The difficulty is reflected in incident resolution times. The mean time to resolve a network incident is 88 hours, while the median is 12.5 hours. The mean is skewed by some organizations that take a week or longer to resolve issues, reflecting the degree of complexity. That gap between mean and median suggests a long tail of hard-to-solve problems that consume disproportionate time and resources.
There is a real opportunity to do a better job by having AI help with analysis and automate the root-cause process. One example is when an organization tends to take the same corrective action every time a situation occurs. If a team has done it 14 times, it makes sense to automate it, but the system should still tell the team it did so. Network professionals may also take a page from their security counterparts. With a similar problem of too many issues to tend to, security pros are increasingly automating the response, even if that means shutting down a resource. The thinking is that the potential losses from inaction are greater than the potential impact to the business. Network teams are moving their way toward being more comfortable with that approach.
The Solution: Another Single Pane of Glass With an AI Twist
The networking vendor is proposing a new cloud control platform as a solution. It is intended to provide a unified view and management plane for networking, security, compute, observability, and collaboration solutions. The platform also applies agentic AI to diagnose and resolve issues, including those that cross domains. It is, yet again, the proverbial single pane of glass, this time with an AI twist.
It may seem somewhat ironic that a vendor that sells the networking gear that has become too complex to manage is now also selling the solution intended to address that complexity. Industry analysts acknowledge that network vendors have been complicit in creating this complexity, but they also say it is good to see them simplify things now. The network is being used in so many more ways than it ever has before. It is more complex for a number of reasons, including that networks support orders of magnitude more devices and connect to pretty much everything.
For those who have watched networks evolve and attempts to automate operations, the goal has been very difficult to achieve. The difference now is AI. With agentic capabilities, strict guardrails, and a willingness to let machines handle routine actions, network operations may finally move toward a model where humans set policy, approve exceptions, and focus on higher-value work. The remaining question is not whether AI will take on operational responsibility, but how quickly organizations can build the trust, controls, and workflows needed to make that responsibility safe and useful.
Source: Network World News