Connect
Ask IT leaders how mature their AI deployment is, and the number keeps shrinking.
Six months ago, 40% said they had it handled. Today it is 23%. Nothing broke to cause that. AI did not slow down. It sped up. Agents took on more access and more autonomous action, and the deeper those agents went into production, the more IT leaders looked around and realized they could not fully see what they had set loose.
That shrinking number is the whole story in itself. It is not a maturity problem. It is a dawning realization. Teams are handing AI more responsibility every quarter while quietly losing track of what it can touch, what it can do, and who is accountable when it goes wrong. The confidence of IT teams is not falling because the work is going badly. It is falling because awareness is rising. The people closest to the deployment are the first to see the size of what they have taken on.
That is the tension our Q3 2026 IT Trends Report brings to light: AI access and autonomy are scaling faster than anyone’s ability to govern them.
To map that gap, we surveyed 800 IT leaders across the U.S. and U.K. The picture that came back is a market still pushing AI forward hard while quietly doubting it can keep the whole thing secure. Deploying the agents was never the hard part. Controlling them is. And most teams know they are behind.
The findings below are not four separate problems. They are four filters on the same one. And they trace a single fault line: the widening distance between what AI can now do inside an organization and what IT can actually govern.
The AI Maturity Mirage
Organizations want more AI. They trust themselves with it less. 84% plan to expand AI use in IT operations over the next six to 24 months, yet only 23% call themselves mature at deploying it, down from 40% half a year ago.
That drop is not a sign teams are getting worse at AI. It is a sign they are getting honest about how quick-moving and deeply technical AI security has become. The more a team rolls AI out, the more it discovers how much sits outside its line of sight.
Confidence does not fade because the work is going badly. It fades because the teams furthest along are the first to grasp the real scope of the job. A shrinking maturity number, in other words, is often what competence looks like on the way up.
The Rise of Non-Human Identities
AI agents are piling into environments at a rapid pace. One in three organizations now runs six or more non-human identities (NHIs) for every human user. Only 21% have any governance in place for them.
That means most IT teams are managing a swarm of autonomous systems with no clear owner, no access review, and no clean way to pull one offline when it misbehaves. Every one of those identities is a way in. Almost none of them are being watched.
The Productivity Reality Check
AI is saving time and creating new work in the same breath. More than 90% of IT leaders still report productivity gains, but the people who are calling AI a major time-saver slipped from 56% six months ago to 45% today.
Some of the time AI hands back is quietly reabsorbed by the work of supervising it. Verifying outputs, securing integrations, cleaning up where the agent got it wrong. The gains are real. So is the new workload nobody budgeted for. What is really happening is a shift in the job itself: less time spent doing the work, more spent overseeing the systems that now do it. That is progress, but it is not free.
The Cost of Fragmented IT
Stacking AI on top of disconnected systems is a security problem. Organizations juggle an average of 6.9 separate tools to run core IT, and 96% say IT unification directly shapes whether they can scale AI safely.
Every extra tool is one more place for identity, access, and oversight to slip through. The more scattered the environment, the less ability you have to govern what the agents inside it are actually doing.
What This Means for Your Organization
These findings paint a clear picture. Access is outrunning accountability. Security has become the number one barrier to scaling AI agents. And most organizations are still trying to figure out how to get the control and visibility they need to really scale their agentic AI efforts.
Read the full Q3 2026 IT Trends Report to see exactly where AI governance gaps are opening, what sets the teams pulling ahead apart from the ones falling behind, and how to build the identity and operational foundation to scale AI without losing your grip on it. Download your copy to see where your agentic AI practices stand.
