The state of AI in healthcare: adoption, ROI, and workforce

Executive brief: Our research uncovers the domains generating the strongest ROI, the practices that distinguish top-performing organizations, and why disciplined governance, measurement, and portfolio management matter more than simply deploying more technology.

Major findings from the survey

1. AI value has moved from theoretical to tangible

Nearly two-thirds of AI deployments deliver returns of 2x or greater, with a median ROI of 2.5x and a mean of 3.2x. Although only 41% of deployments report a direct financial impact, nearly all report some form of benefit, including cost avoidance, productivity gains, increased throughput, improved quality, and a better patient experience.

The survey highlights that AI value is no longer limited to isolated pilots or future projections. For health system leaders, the conversation is shifting from “Should we invest?” to “Where are returns strongest, and how do we scale what works?”

2. AI’s biggest opportunities may be hiding in unexpected places

Pharmacy, supply chain, and utilization management all report median returns of 3.0x despite adoption rates below 50%. Meanwhile, several of the most widely adopted domains report more modest returns, including clinical care at 2.1x and patient access at 2.0x.

This doesn’t mean that organizations should pull back from common applications, which may deliver important quality, productivity, or patient-experience benefits. Instead, the data points to the value of looking beyond the most visible use cases.

3. The highest-performing organizations know when to invest and when to walk away

Compared with peers, ROI Leaders add more tools, scale more solutions, and retire underperforming technologies more frequently. Over the previous 12 months, they added a median of six tools, expanded eight and retired two, compared with four added, five expanded and one retired among Emerging ROI organizations.

The results underscore that stronger returns are associated with active portfolio management rather than simply accumulating more technology. These organizations are also more likely to use formal criteria for decommissioning low-value tools and to track benefits after implementation.

4. AI changes work more often than it eliminates jobs

Among ROI Leaders, the most common workforce impacts are role redesign (59%), staff redeployment to higher-value work (56%), and reduced overtime without reducing headcount (53%). By comparison, 21% report reducing headcount through attrition and 17% through layoffs.

This reinforces the idea that AI value often comes from reorganizing work and making more effective use of limited staff capacity, not simply removing positions. When AI frees employees’ time, organizations report redirecting that capacity toward training, new initiatives, and quality or safety improvement, creating an opportunity to reinvest productivity gains in continued operational improvement.

5. Operating maturity is emerging as the true competitive advantage

The organizations achieving the greatest value from AI are not distinguished by technology alone. ROI Leaders deploy AI across an average of 8.6 domains, compared with 7.5 among Emerging ROI organizations, and 42% operate mature, scaled portfolios vs. 20% of their peers. They’re also more likely to have enterprise-wide governance, track post-implementation benefits, and apply formal criteria for retiring low-value tools. More consistent measurement helps connect AI performance to decisions about what should be scaled, funded, renegotiated, or stopped.

Taken together, the findings suggest that sustainable value depends on the operating model surrounding AI, including governance, accountability, measurement, and disciplined portfolio management.