Will Health Technology Meet the Affordability Moment?

We hosted our third PHTI Summit in New York City on June 4th, bringing together a diverse group of healthcare leaders—employers, health plans, hospital systems, clinicians, and investors—for an honest conversation about using health technology to address affordability.

The sessions explored critical issues, including the clinical and economic potential of artificial intelligence, expanding employer purchasing power, and accelerating the adoption of outcomes-based payment models. The same tension ran through every session: in today’s system, it remains easier to raise prices than to raise efficiency.

Every panel circled back to the same central idea: we have a generational opportunity to leverage the power of purchasers to realize better patient outcomes at lower costs—instead of accepting the status quo of raising prices to make room for new layers of technology.

The story is familiar: consolidated systems hold pricing power, and payers that capture efficiency book it as margin rather than passing it through. Raising prices is a move every incumbent already knows how to make; raising efficiency is a capability almost no one is paid to build. And every actor’s margin comes out of the same limited stack of premium dollars. Technology deployed into that environment does not bend the cost curve. As one participant said, technology—particularly AI—dropped into a dysfunctional system is an accelerant and an amplifier: it gives each actor a faster way to do what it’s already doing. But as another participant observed, that is not a technology problem. It’s a structural one—and we haven’t redesigned the payment mechanisms and operating models the technology sits on. Until we change what we pay for, every new capability will simply reinforce the system we already have.

One path to ensure that investments in technology deliver better health at lower costs is outcomes-based payments, which shift from paying for activity to paying for results. Three years ago, paying for outcomes was aspirational. Today, nearly half of digital health purchasers use performance-based contracts—and vendors are putting their payments at risk to prove their solutions work. Now Medicare is adopting the same approach. CMS’s ACCESS model is testing outcomes-aligned payments in traditional Medicare, expanding access to technology-supported chronic care for millions of beneficiaries.

These new payment models purposefully shift away from paying for vendor platforms to paying directly for measurable health improvements, such as clinically verified drops in blood pressure and hemoglobin A1c levels. The model is deliberately prescriptive about measurement—readings must arrive as automated, time-stamped entries transmitted directly from authorized medical devices—to prevent gaming or data manipulation. And where most outcome-based contracts today chase small, defensive savings targets, ACCESS deploys outcome-based payment at scale, building shared infrastructure that commercial purchasers can adopt and adapt to move away from activity-based fee schedules toward accountability for total cost of care.

Yet, the audience had a concern: how do we drive better engagement with digital health solutions? Outcome-based payment only pays off if people actually use the care. One purchaser described ending vendor relationships that could not sustain engagement above 10% of the eligible population; an employer countered that real engagement is never vendor-led—it comes from the employer putting the benefit in front of employees again and again. As one participant put it, engagement is the blockbuster drug of our lifetime. And no technology has more to gain from a shift to outcome-based payments—or more to lose without it—than AI.

AI could finally lower healthcare costs. But will it? Our own reports have already found that administrative AI is driving up costs in prior authorization and billing because it is being deployed into payment systems that reward volume. As currently deployed, administrative AI has sparked a costly “bot war,” becoming an accelerant to systemic dysfunction rather than a solution. Payers use automated systems to reject or delay care, while health systems deploy opposing bots to automatically fight back and maximize billing intensity.

Clinical AI applications hold enormous promise but also face important barriers including a lack of viable payment pathways, regulatory pathways, and insufficient oversight processes to monitor models that evolve over time. Summit participants also flagged that these tools may not always be cost-reducing, especially when you consider the millions of dollars in annual implementation and ongoing token costs. They described enterprise AI contracts running into eight figures a year and engineers consuming six figures a year in token costs alone.

There is an opportunity for AI to change clinical care, particularly for conditions where access is low, outcomes are poor, and current approaches are costly. However, getting there demands something of everyone: developers to move from assistive tools to autonomous care, health system leaders and patients to trust that clinical AI works, and regulators to clarify coverage and payment policies that encourage safe deployment and a commitment to net savings. Until we address those barriers, efforts to advance clinical AI will continue to lag, with emphasis on assistive AI that fits into existing workflows and payments. The growth of payment systems that focus on outcomes, instead of process and activity, is a meaningful bridge to the future. But we also need to develop ways to pay for results-based clinical AI that can be delivered at scale.

The discussion kept returning to one question: who actually buys AI designed to reduce cost rather than increase billing? For administrative AI, the buyers exist—but health systems and commercial payers deploying these tools are structured to capture efficiency gains as revenue, not pass them through as savings. For clinical AI, we need a national shift to purchasing clinical outcomes. The CMS ACCESS program is a signal that the nation’s largest purchaser, traditional Medicare, believes there is a scalable path to purchasing chronic condition outcomes.

Employers and government systems are the entities with the most to gain from deflationary AI, and they are starting to model the purchasing behavior we need to realize these gains. For AI to be truly deflationary, the payment models that reward it must exist—and until they do, powerful AI tools will continue to sit on the shelf. The window to get this right is now. If we act, AI could meaningfully lower costs and improve outcomes at scale. If we wait, we will miss the opportunity.

Stakeholders across the sector recognize the significance of this moment—particularly employers. They cover more than 165 million people and spend nearly $1 trillion annually on healthcare. For three consecutive years, their premiums have risen more than 6%—the first time that’s happened in two decades—and early signs point to an even steeper climb this year. Employers absorb these increases and share the pain with their workers through lower wages, higher premiums, deductibles, and out-of-pocket costs.

Employers hold tremendous purchasing power, but it has gone largely unused—not for lack of motivation, but for lack of information: most have no way of knowing whether the prices they pay are competitive, whether the providers they cover are delivering high-quality care, or whether their vendors are effectively negotiating on their behalf.

Direct contracting is one solution that jumbo employers are adopting to reshape market dynamics. One purchaser explained how they had reviewed their claims data and identified that they were paying $1,300 for a colonoscopy at one site and $12,000 at another. That unjustified price variation led to years of data analysis, integrating hospital price transparency data, so that when a health system came to the table, the purchaser could do what many employers cannot: reprice the entire offer against its own claims to verify the value was real. The contract closed in months, on terms only a data-equipped buyer could write—line-item price caps so no single price can rise more than 5% a year, and free primary care visits to encourage employees to see their doctors.

What makes this example matter is that it was built to be replicated. On the health system side, the arrangement runs on plug-and-play infrastructure—an off-the-shelf model that can be deployed rapidly across other employers rather than rebuilt from scratch. On the member side, engagement was treated as a deliverable, not a hope: a dedicated call center helps members schedule care, manage referrals, and navigate the system, with outreach timed to members’ work hours and offered in their languages. That navigation layer is both an engagement strategy and an early signal of outcomes to come.

The lesson from that project and the panel as a whole was the same: when employers have clear, independent evidence to guide their purchasing decisions—analysis with no financial ties to health plans, health systems, or benefits consultants—they can make smarter decisions, demand better outcomes from vendors, and bend the cost curve. The potential for cost savings if more purchasers have access to that kind of data is enormous. As one participant put it, the goal isn’t to spend less than your peers—it’s to move the whole market.

The health technology sector is rapidly evolving, but if we are willing to be clear-eyed about what is working and what is not, health technology can deliver on its promise of better outcomes at lower costs. PHTI will continue pushing deeper into these questions. In the weeks and months ahead, that means new analysis of AI’s clinical and economic impacts, of the ACCESS model and outcome-based payments, and new assessments of the technologies targeting the chronic conditions that drive so much of our healthcare spending. We will continue probing these hard questions with key leaders from across the sector. And will continue our core work of accelerating the adoption of technologies and tools that will help us realize a better, more affordable, and more efficient healthcare system for every American.

We’re grateful to the employers, health plans, health systems, policymakers, and investors who joined us in NYC for these critical conversations.