A Netrio commissioned survey of 401 U.S. IT leaders at companies with 200 to 5,000 employees finds deployment is outrunning the policies, visibility, and security controls that would make AI a managed program rather than a scattered experiment.
Eight in ten mid-market companies have put AI into production. Barely a quarter run it like a real program. That gap, between widespread deployment and the governance infrastructure that would make it sustainable, is the central finding of a Netrio-commissioned survey released this week, and it lands in the middle of a market where the distance between AI ambition and AI maturity is coming into sharper focus.
The survey, fielded by Censuswide for the managed service provider Netrio, polled 401 U.S. IT leaders at organizations with 200 to 5,000 employees. Eighty-two percent of respondents said AI is already in production or in widespread use somewhere in their organization. Only 26 percent said AI is scaled and governed across the enterprise, a gap Netrio's report frames as the divide between adoption and operational maturity.
The sponsorship matters. Netrio is a global managed service provider headquartered in McKinney, Texas, and the report is built to highlight exactly the services Netrio sells: implementation, governance, and security. Censuswide is a third-party research firm that ran the survey on Netrio's behalf. The underlying report PDF, full methodology, and question wording were not in hand at the time of writing, so the headline numbers should be read as directional, not definitive, and as a self-reported perception metric rather than an audited deployment measurement.
The governance picture inside the survey is more revealing than the adoption number. Just 42 percent of respondents said their organization has an enforced AI policy. A separate 53 percent said they have full visibility into which AI tools are being used across the business. The combination is the kind of operational problem a non-beat reader can picture: tools running in production, but no policy enforcing the rules and no inventory showing where those tools live.
When asked about the biggest barriers to scaling AI, respondents pointed to security, privacy, and compliance first, at 19 percent, with data readiness at 17 percent and integration complexity at 16 percent. Those are the practical problems that follow from putting a model into production without a program around it: where the data sits, how it moves, who can use it, and whether the security team can see any of it.
Forward-looking sentiment runs hot. Ninety-six percent of respondents said they are confident AI will deliver measurable ROI within 24 months, and 88 percent said their organization plans to spend at least $100,000 on AI in the next year. Vendor surveys reliably produce optimism that maps to the vendor's sales motion, and these are the numbers to weigh with that in mind.
The deployment-versus-governance gap is the story the data actually supports. Adoption is no longer the constraint for most mid-market IT leaders. The constraint is the slower work of writing the policies, building the inventory, and integrating the security and compliance checks that turn scattered AI experiments into a program the business can actually run.