There’s a gap in health system AI strategy that doesn’t get enough attention, and it isn’t about governance frameworks or pilot programs. It’s about how much of the IT team’s time is disappearing into vendor management — and how far the industry has drifted from how most technology leaders say they want to operate.
Our 2026 CIO Research Report found that 72% of health system technology leaders would prefer to work with one comprehensive AI partner over managing multiple specialized vendors. Only 11% currently do.
The hidden cost of fragmentation
More than half of respondents — 51% — report spending between 11% and 25% of IT bandwidth on vendor management, integrations, and implementations alone. 17% report that figure climbing as high as 50%. When IT teams are running that kind of overhead just to keep existing tools working, there’s very little capacity left to evaluate new solutions, drive adoption, or measure outcomes.
“One of the biggest bottlenecks is procurement — RFI, RFP, questionnaires, AI governance — that process alone can take several months,” as one CMIO at a Southwestern health system described it.
Many health systems pursued vendor diversification in order to get best-in-class capabilities. The result, for most, is a fragmented stack that doesn’t perform well and is cumbersome to manage.
What health systems actually want from a partner
When we asked what matters most in an AI partnership, the priorities were clear. 63% cited proven clinical outcomes with strong references as a top criterion. 59% want guaranteed ROI or a risk-sharing pricing model. 49% prioritize financial stability and long-term visibility.
The underlying message across all three is the same: health systems want partners who are accountable for outcomes, not just for delivery.
“You may have one specific, compelling solution; that’s not what we’re looking for,” said Matthew Anderson, MD, CMIO at HonorHealth. “We’re looking for things that have enterprise impact — a partnership that is collaborative and long term.”
Morgan Jeffries, MD, Medical Director for AI at Geisinger, described the direction simply: “We’re trying to partner with as few groups as possible.”
The workforce dimension
Vendor sprawl doesn’t just strain IT resources. It burns out clinical staff. 87% of respondents identified “tech fatigue” — staff exhaustion from learning new systems before seeing tangible value — as a barrier to AI adoption. Every new point solution added to the stack is another change management effort, another training cycle, another window for adoption to stall.
The health systems making real progress are doing so in part by consolidating around fewer, deeper partnerships that don’t just deploy technology but also take responsibility for driving its adoption and measuring its results.
How to close the gap
The math on vendor sprawl is not sustainable, and the survey data suggests most technology leaders already know it. The question is how to move from a fragmented stack built up over years of individual purchasing decisions to the consolidated model that 72% say they’d prefer.
Our report details what that consolidation shift looks like in practice — including the criteria CIOs are using to decide which partners to deepen and which relationships to exit.
