Just 14% of healthcare CIOs say their organization has a comprehensive, well-defined AI strategy, according to Qventus survey data and interviews with CIOs, CMIOs and senior IT leaders at medium and large U.S. health systems.
While enthusiasm for AI remains high, two-thirds (66%) of respondents say they are still developing their strategy, and 20% describe their approach as "limited or fragmented."
The report paints a picture of healthcare IT leaders under pressure to evolve rapidly, with CIOs now responsible for strategic decisions around AI use cases, governance structures, and whether to build solutions internally or buy them from vendors.
More than half of the CIOs surveyed say their primary role is leading strategy development and execution, marking a significant shift from traditional IT responsibilities.
To move AI from experimental to a core part of a health system, the most important thing to recognize is AI is a tool in the toolbox, Mudit Garg, Qventus CEO and cofounder, told HealthcareFinanceNews. Qventus provides AI-powered software to automate healthcare operations.
"It's a really impressive tool with broad implications, but to be effective, we can't do AI for AI's sake," he said. "Our work needs to be grounded in outcomes and ROI – patient benefit, financial benefit and staff benefit."
Despite this push, just 2% of CIOs said they believe their EHR system is mature in its AI capabilities; 60% said it's still early days.
Respondents cited a lack of internal resources or expertise to develop in-house AI tools, with most relying on EHR vendors due to workflow familiarity, even though vendor-provided solutions often fall short of expectations.
Nearly six in ten CIOs said their EHR's AI features delivered ROI as expected, but 40% said the ROI has been underwhelming.
When evaluating whether to build or buy AI tools, CIOs rank cost comparison, return on investment size, and speed to value as the top considerations, while customization and time to implementation ranked lower.
Among those who opt to build, 63% said they measure internal AI development costs primarily through team labor and EHR integration.
When weighing third-party vendors versus EHR-based AI, CIOs cited functionality, integration capabilities and cost as the most important factors.
Garg advised healthcare organizations to identify challenges that are holding the organization back – for example, misutilization of OR time or increasing ED boarding times – and then see if AI can be a meaningful accelerant.
"This creates a cultural, financial and organizational juggernaut of success leading to more willingness to use AI more broadly," he explained.
The report also found that more than half of surveyed organizations now have a formal AI governance committee – typically composed of IT leaders (94%), clinicians (78%), CFOs (62%) and regulatory experts (62%) – tasked with overseeing safe, effective deployment of AI tools.
Clinician frustration also remains a significant barrier to adoption. CIOs reported that the top complaints about AI tools from clinical staff include poor integration between systems (34%), lack of clinician input in design (24%) and frequent changes to the tech stack (16%).
Garg pointed out AI is only successful when surrounded by deep understanding of data, workflow, user behaviors and goals.
"IT leaders must help the organization focus deeply on the 'job to be done' by the AI to make sure it is successful," he said.
He noted hospitals have been burned by tools that sound great in theory but suffer from low adoption and a failure to integrate with, and then improve, existing workflows.
"From identifying the problem through implementation, you need to have everyone at the table: leaders, clinicians, IT folks and vendors," Garg said.