Everyone agrees (or has at some point) that value-based care is the future. Payers and providers nod along in the same meetings, shake hands, and then go back to their separate corners to do... whatever they were already doing.

AI in healthcare is the other major point of discussion in most boardrooms these days. The Menlo Ventures 2025 State of AI in Healthcare report suggests that healthcare is deploying AI more than 2x as quickly as the broader economy. But too often it stays there, in discussion.

But here's what gets interesting. These two often-struggling initiatives have a symbiotic relationship, and can benefit from each other to succeed. In this article we’ll walk through why that is and how to do it.

Three Reasons Value-Based Care Keeps Stalling (And How AI Can Help)

Value-based care has a data problem. Actually, it has several data problems.

The first is fragmentation. VBC requires seeing the whole patient. Their clinical history, social circumstances, what happens between visits. But that information lives in dozens of different systems that don't talk to each other.

GenAI with natural language processing can act as a de-facto system integrator here. Not by solving the interoperability problem, but by synthesizing information from clinical notes, claims data, social determinants assessments, and other sources into something coherent. Kearney's analysis makes the case that this "soft integration" may be more practical than waiting for true technical interoperability.

The second problem is predictability. Providers are reluctant to take on risk-based contracts, because they can't reliably predict costs for a patient population.

AI changes this calculus by identifying which patients are likely to deteriorate, which interventions are most cost-effective, and where resources should be concentrated.

The third problem is measuring what matters. VBC is supposed to reward patient-centered outcomes—quality of life, functional status, patient satisfaction. But collecting this data at scale has always been impractical.

AI makes it feasible through automated collection, natural language processing of patient feedback, and personalized digital interfaces that meet patients where they are.

The Symbiotic Relationship

Value-based care gives AI a clear business case.

When contracts reward outcomes, investments in outcome-improving technology make obvious financial sense.

Meanwhile, AI gives value-based care the ability to scale. Population health management, risk stratification, care gap identification, outcome prediction… all of these are impossible to do well manually across large patient populations.

Neither works particularly well alone. VBC without AI is aspirational but impractical. AI without VBC is technically impressive but financially unjustifiable.

Together, they create something neither can achieve independently: a sustainable model where doing the right thing for patients is also the right thing for the business.

Where This Is Actually Working

This isn’t theoretical. The numbers from real implementations are striking.

• Johns Hopkins leveraged AI to achieve a 20% reduction in sepsis mortality by detecting the condition an average of 6 hours earlier than traditional methods.
• Geisinger Health System used AI to reduce the time from nodule discovery to follow-up vision from 112 days to just 8, with zero missed malignancies. For colorectal cancer, a predictive algorithm identified high-risk patients, increasing the cancer diagnosis rate from 1 in 154 to 1 in 13 for the targeted group.
• Post-surgery chatbots have dropped readmission rates from 8.3% to essentially zero in some orthopedic programs. One study found ED visits fell from 8% to 0.9%—a 90% reduction.
• Cleveland Clinic’s Ambient AI Scribe saves providers an average of 14 minutes per day in the EHR and reduces time spent on notes by 2 minutes per appointment.

These aren't research projects. They're production systems generating measurable improvements in both clinical outcomes and financial performance.

How To Make It Happen

Based on what we're seeing with healthcare clients, a few patterns seem to separate the organizations making progress from those stuck in pilots.

• Start with operational wins, but don't stay there. Administrative automation builds confidence and frees up resources. Prior authorization workflow improvements, scheduling optimization, documentation assistance… these are proven areas that can create quick wins. But you don’t stop there. You treat them as stepping stones to clinical applications where the ROI is exponentially higher.
• Governance before scale. The organizations that jump straight to enterprise-wide AI deployment usually regret it. Building governance structures, validating models for bias, and clarifying decision rights takes time, but prevents the trust failures that kill adoption.
• Invest in the workforce first. There's a direct line between employee satisfaction and patient satisfaction. Using AI to reduce administrative burden (letting clinicians focus on patients instead of paperwork) improves both employee experience and VBC quality metrics like CAHPS scores.
• Pick conditions where clinical and financial ROI align. Sepsis, heart failure, stroke, post-surgical complications… these are areas where better outcomes directly translate to better financial performance under value-based contracts. The alignment isn't accidental; it's the whole point.

What This Really Means

For years, healthcare has had a frustrating gap between what we know works and what we can actually do at scale. Value-based care pointed in the right direction, but couldn't get there. AI had the horsepower, but no clear destination.

The opportunity now is to close that gap. Not necessarily perfectly, but meaningfully. To build systems where preventing a readmission isn't just the right thing to do, it's also the smart thing to do. Where the technology serves the mission instead of the other way around.

That's not a small thing. And it's finally within reach.

‍