More than half of health system and health plan executives say AI is an immediate priority, and 73% are increasing their investments in the technology, a new C-suite survey finds.
Many healthcare organizations are moving past early pilot successes to enterprise scaled solutions, but are balancing AI enthusiasm against pragmatism, according to the survey from Define Ventures of C-suite and senior executives leaders from more than 60 providers and payers.
Define Ventures, a venture capital firm focused on early-stage health tech companies, conducted surveys and meetings with executives from 10 of the top 20 providers and three of the top 10 payers to check the industry's pulse on AI adoption and investment. The survey took place from August through early November.
Providers and payers are confident about AI's immediate opportunity to improve the patient and clinician experience over the next two years, as indicated by 58% of survey respondents. But, they are less bullish on AI's ability to bend the overall healthcare cost curve downward or to impact the quality and access to care over the next two years (33%).
"I think some of the top-level aspects that we see is there is a lot of ambition and optimism. For providers, AI is not a nice to have. They are saying, 'AI is essential for us to continue to operate in our healthcare system'," Lynne Chou O'Keefe, founder and managing partner at Define Ventures, told Fierce Healthcare.
Almost three-fourths (73%) said their organizations have set up an AI governance structure to set priorities and plan for AI's future and 22% said they are establishing a governance committee. Among the organizations with established governance structures, the primary focus areas include identifying and prioritizing use cases (91%), establishing ethics and safety guidelines (87%) and setting data policies (84%).
These committees play a critical role in aligning AI initiatives with organizational values.
Most providers and payers are focused on AI projects that offer "quick wins" in operational efficiency, which can be essential for achieving buy-in, building creditability and establishing a sustainable foundation for more transformative applications down the road, the report found.
"With AI, you can't sit on the sidelines. This space is moving too fast; you will be so far behind. So I think a lot of our partners understand pragmatically that you need to jump in. They're obviously taking the use cases that reflect 'soft ROI' metrics that are very acute and can be felt, which I think is very smart," O'Keefe said. "I think they're just being really smart and thoughtful of where to take those use cases for early, near-term wins, while laying the foundation for the midterm to long-term value drivers."
With this strategy in mind, 83% of health system leaders say ambient scribing, which automates clinical documentation, has become a top priority. These tools are seen as essential for clinician retention and recruitment and can build momentum for future AI initiatives, according to the report.
Other priority use cases health system executives identified include administrative AI for financial management (59%), clinical AI for disease screening (39%), patient communication (31%), supply chain management (21%), clinical decision support (21%), research and analytics (17%) and payment verification (17%).
"While the potential for AI in healthcare is immense, identifying where to invest is critical," John Halamka, M.D., Mayo Clinic Platform President, said in a statement included in the report. "Our team's focus on eight key areas enables us to align AI capabilities with organizational priorities. It's not just about the technology, it's about creating a healthcare system that works smarter for both patients and clinicians."
There's ongoing discussion about defining ROI for AI initiatives. Many health system leaders said they are not focused on financial ROI today, but are learning and iterating, according to the report.
"Defining ROI is anything but straightforward," said Richard Milani, M.D., Sutter Health Chief Clinical Innovation Officer, in the report. "Take ambient scribing tools, for example, Should we focus on the number of additional patients seen by our clinicians, or by the number of doctors getting home in time for dinner with their families?"
"Yes, we're focused on operational efficiencies and top business priorities. But I'm just as focused on knowing our organization is doing the right thing for our clinician workforce," Milani said.
The majority of payers (68%) see the opportunity to improve the members' experience but largely through operational efficiency improvements such as call center optimization. Other use cases cited by payer executives include case planning and care planning (63%), utilization and prior authorization (56%), population health (44%), finance management (25%) and billings and claims (25%).
As healthcare organizations develop their AI strategies they are confronted with the classic "build or buy" decision. Currently, 72% of payers and providers are leaning on external partners for application solutions, the survey found. Still, about half (51%) are also building some AI solutions in house, according to the survey.
Looking at the technology stack, with the compute layer (cloud, GPU) and the LLM layer, most healthcare organizations are relying on external vendors. But when it comes to data infrastructure, most are investing heavily in internal solutions for data aggregation.
When it comes to the application layer, currently, 72% of payers and providers are leaning on external partners for application solutions, but this could change as the technology evolves, the report notes. "This landscape is rapidly evolving as AI technology becomes more accessible and commoditized. Capabilities that seemed complex months ago are increasingly becoming more commonplace," the report said.
Organizations also want seamless integration with existing systems and workflows, particularly with providers and EHR platforms, and that also is shaping the "build vs. buy" decision.
Providers and payers also face integration challenges, and 64% of C-suite leaders said establishing a clear ROI as one of the biggest hurdles. Forty percent said their technical teams' bandwidth was a challenge, 38% of executives said the integration process itself is a major challenge, 24% identified data usage, ownership and privacy issues, 18% pointed to change management challenges and low adoption.
When evaluating AI vendors, providers cited ROI for the chosen use case (45%), seamless workflow integration (38%) and data security and compliance (38%) as top considerations. For payers, 86% cited data security and compliance, followed by data access and ownership, ROI for specific use cases.