Mature governance is key to scaling AI in healthcare

The state of AI in healthcare 

Sharing his perspective on the current trajectory of AI in the healthcare sector, Mysore notes that while more than 90% of U.S. hospitals now use some form of electronic health records (EHRs), this development should be seen as a digital milestone rather than a transformative one. 

Referring to an industry study, Mysore says that AI has the potential to unlock $200 to $300 billion annually in U.S. healthcare through: 

• Improved care outcomes
• More efficient operations
• Reduction in waste 

However, Mysore noted that only about 15% of that potential is being realized today. That lag is not due to limitations in AI models or technology. Instead, he points out the systemic challenges such as complex workflows, inconsistent clinical processes, widely varying practices, and AI being forced into systems that aren’t ready for it. 

Another critical trend Mysore identifies is the proliferation of AI pilots without meaningful scale. 

“Everywhere you look, there are a lot of pilots being done, which is great, but very few of them are actually scaled to deployments.” 

While many healthcare systems are exploring AI, Mysore said many are stalled at the pilot stage due to: 

• Siloed data 
• Non-standard processes 
• Limited return on investment 
• Lack of trust 

AI governance as a catalyst 

Delving further, Mysore stresses that mature AI governance is key to driving successful, scaled deployment. 

 “Organizations with mature governance—AI governance—are about 2.3 times more likely to scale AI and deploy successful AI initiatives,” Mysore said, noting he has confirmed this in his own experience at Sutter Health. “Without clear guardrails, AI doesn’t get traction,” he adds. 

Mysore said he remains optimistic but clear-eyed about what’s needed to move forward. 

“If I zoom out now, the opportunity with AI is massive. But to move from the promise of AI to real progress and outcomes, we need better workflows. We need better integration of AI into some of the frontline care; we need trust.” 

Balancing personalized care with operational efficiency 

Reflecting on the evolving demands of modern healthcare, Mysore highlighted the central challenge of delivering deeply personalized, patient-centered care while also managing the rising pressures of operational efficiency. 

Furthermore, he emphasized that part of addressing this tension involves redefining how value is measured in healthcare. “Hospital operating costs have surged over 20% in the last few years. At the same time, labor is getting expensive. Supply chain volatility and inflation are driving the costs up. This is the upside. And then, the reimbursements ultimately stay flat.” 

As a result, health systems are being asked to do more with less — without compromising quality of care. “This is where the ROI conversation gets critical,” says Mysore. 

A new lens on ROI in the age of AI 

Traditionally, return on investment (ROI) in healthcare has focused narrowly on cost savings or revenue enhancement. Mysore suggests that this view is outdated in the context of AI and digital transformation. “In the world of AI and digital tools, we need to go further,” he said. 

Mysore proposed a broader framework, one that includes both input and output metrics, which ultimately connect to patient outcomes.  

“Input metrics are things like time saved, reduction in repetitive tasks, and so on,” he said. “Output metrics are really what the organization is striving towards, which is reduced burnout, higher clinician satisfaction, improved patient experience, and so on.”