AI in Healthcare — From Noise to Meaningful Transformation

“Healthcare's AI opportunity isn't limited by technical capability but the willingness to reimagine.”

‍

OpenAI co-founder & president Greg Brockman’s perspective at our recent AI Summit captured the heart of this moment. Healthcare has the data, the need, and now the technology to transform from the ground up. The question is no longer if AI matters — it’s how deeply we’re willing to let it reshape the system.

‍

Healthcare has experienced meaningful waves of technological change — from the digitization brought on by electronic health records to the expanded access enabled by virtual care. Each brought progress, though often uneven and slower than hoped. AI builds on those advancements, but represents a more fundamental shift: a convergence of systemic pressures, technical breakthroughs, and market readiness that makes large-scale transformation not just possible, but inevitable.

‍

• Systemic pressures. Rising costs, workforce shortages, and widening quality gaps are forcing the healthcare system to rethink how care is delivered. By 2028, the U.S. is projected to face a shortage of more than 100,000 healthcare workers — a gap too large to close with traditional models. At the same time, demographic shifts are intensifying demand: by 2030, one in five Americans — over 71 million people — will be 65 or older, placing unprecedented strain on already limited resources.

‍

• Technological breakthroughs. Advances in compute and model architectures have fundamentally changed what’s possible in healthcare AI. The cost of training large models has fallen by orders of magnitude, while multimodal approaches now allow for seamless integration of clinical text, images, and even voice. This leap isn’t theoretical, it’s already reshaping use cases across the ecosystem.

‍

• Market readiness. For the first time, enterprise healthcare buyers are eager — not reluctant — to engage with startups. According to our research with C-suite executives,  53% of payers and providers and 70% of pharma companies said AI is an immediate or urgent priority, spurring the creation of governance councils to manage adoption.

‍

Define’s Perspective on the AI Opportunity

‍

Define Ventures was founded on a simple belief: the future of healthcare will be shaped by those who combine deep knowledge of the healthcare ecosystem with a technology-first mindset. In healthcare, true progress requires partnership — innovators and incumbents must work together to create lasting change.

‍

We built Define Ventures around this vision, growing into one of the largest funds dedicated to early-stage health tech innovation, and we’ve been preparing for this moment since our founding. Our work with early AI companies and our collaboration with industry leaders give us a unique vantage point into both the potential and the practical realities of deploying AI in healthcare. We’re also proud to have world-class AI advisors Dr. Nigam Shah, Chief Data Scientist at Stanford Health Care, who focuses on responsible translation of AI into clinical practice, and Dr. Jia Li, former Head of Cloud AI at Google, who brings expertise in scaling infrastructure and applying AI responsibly across industries.

‍

As the pace of innovation accelerates, this role of supporting visionaries and incumbents to reimagine healthcare together has never been more crucial. What has become clear to us is that a brilliant algorithm without a clear ROI, seamless workflow integration, or trust from end users will stall. Conversely, solutions that account for enterprise realities can achieve transformative adoption, even if the underlying technology evolves over time.

‍

This perspective doesn’t just shape how we invest — it shapes how we partner. By sitting at the intersection of founders building at the edge of what’s possible and enterprises under pressure to modernize, we see patterns across the ecosystem that few others do. That vantage point allows us to help startups anticipate buyer needs earlier, guide strategics toward transformative partnerships, and continuously refine our own view of where AI will have the greatest impact. 

‍

Where AI Is Creating the Most Value

‍

One way we think about AI’s role in healthcare is through what we call the “House of Healthcare” — an analogy for how value is created across the system. Each layer of the house addresses a different set of problems, but none of them operate in isolation; the true potential comes when the layers connect to form a cohesive structure.

‍

‍

The Front Door: Atomic Personalization

‍

The “front door” is where patients, members, and consumers first engage with the healthcare system — whether through triage, scheduling, care navigation, or member engagement. Historically, this entry point has been impersonal and reactive, built on population-level assumptions: generic symptom checkers, broadcast outreach campaigns, and care pathways based largely on broad clinical groupings like chronic conditions or age brackets. These models often missed the mark and fail to reflect the full context of an individual’s health history, behavior, or preferences.

‍

AI enables a new paradigm: care discovery that is proactive, tailored, and continuous. For example, 68% of payers now rank improving member experience as their top AI use case, signaling a shift away from cost containment alone to precision engagement that drives trust and retention. With atomic personalization, every individual can have their own “door,” shaped dynamically by their risk factors, medical history, behavioral signals, and engagement patterns. Imagine triage flows that anticipate a patient’s likely needs before symptoms escalate, or navigation tools that surface relevant care programs before a member even begins searching.

‍

This level of personalization transforms the front door from a generic intake process into an intelligent system that initiates, guides, and sustains relationships. It not only elevates experience, but also improves downstream outcomes — from adherence to retention to reduced total cost of care.

‍

The Foundation & the Hallways: Data Liquidity & Insight Extraction

‍

Beneath every layer of the house sits the foundation: the unified data layer that enables everything else. Healthcare is notorious for its fragmentation — claims in one system, clinical notes in another, behavioral data scattered across devices and point solutions. For decades, entire teams have focused solely on extracting, cleaning, and reconciling this data through painstaking manual processes. AI can automate much of this work. Foundation models can be used to ingest, normalize, and load disparate data sources at scale, transforming messy inputs into a single usable source of truth. This isn’t just an IT challenge, it’s the prerequisite for every other application of AI in healthcare. 

‍

Once data is unified, the next challenge is making it actionable — translating raw information into insights that guide decisionmaking. We think of this as the hallways of the house: the connective spaces where information flows between rooms, informing movement and coordination. Here, AI excels at surfacing patterns efficiently and scalably — predicting patient risk, flagging gaps in care, or prioritizing clinical workflows. These insights are beginning to power clinical decision support and emerging forms of real-time intelligence, an area we see as forward-looking and poised to become the next major vertical of healthcare innovation. Similarly, in the life sciences, AI will be leveraged to derive insights that impact the discovery, development and distribution of drugs.

‍

Layer Health is at the forefront of this transformation in redefining how healthcare organizations unlock value from unstructured clinical data. Layer’s AI platform automates chart review across a wide range of use cases  — enabling non-technical users to extract reliable insights from complex, longitudinal medical charts for clinical registry abstraction, quality measurement, and care management. Already deployed with leading partners including Intermountain, the American Cancer Society, and Froedtert, Layer’s technology reduces costs and drives new value. For example, Layer’s platform reduced human chart review time at Froedtert by 65% while meeting or exceeding human performance.. The company is led by co-founder & CEO David Sontag, a professor of computer science and AI at MIT whose pioneering research in AI and healthcare has garnered over 20,000 citations across 140 publications. As healthcare’s digital transformation accelerates, this foundational capability positions Layer to become a critical enabler of more efficient, effective care delivery.

‍

But insight extraction isn’t valuable in isolation. The winners in this layer will be those who deliver insights into the moments that matter — embedding recommendations directly into clinician solutions, patient facing applications, or systems of record within pharma. For founders, this is where curiosity about customer workflows becomes a true differentiator: founders who deeply understand inflection and decision making points within workflows can design insights that meaningfully influence outcomes, drive ROI, and pave the way for future platform expansion.

‍

The Rooms: Workflow Automation

‍

Finally, inside the “rooms” is where care is delivered and work is processed, AI is streamlining both clinical and operational workflows — two domains that have historically been siloed but are equally critical to system performance. On the clinical side, AI is reducing the burden of tasks like documentation, chart review, patient recruitment for clinical research, and even patient follow-up, enabling providers to practice at the top of their license and focus on patient care rather than administrative work. On the operational side — the “back office” — AI is transforming functions like medical writing in pharma, prior authorization, scheduling, revenue cycle management, regulatory and compliance capture, and inventory procurement, areas that often determine both financial performance and patient experience. 

‍

Cohere Health, which we incubated with Humana in 2019 and has since expanded to serve multiple payers across the U.S., is demonstrating how companies can evolve within this layer and beyond it. Initially focused on automating prior authorization, Cohere has evolved into a clinical intelligence platform that transforms payer–provider collaboration and redefines how utilization management and care coordination are managed. Its precision clinical insights enable up to 90% of requests to be auto-approved, dramatically reducing provider friction, accelerating time to care, and freeing physicians to focus on patients rather than paperwork. Today, Cohere processes more than 12 million prior authorization requests annually for over 660,000 providers nationwide — driving measurable administrative and clinical efficiencies while enabling deeper collaboration on the most critical cases.

‍

Luminai is redefining how healthcare organizations eliminate manual workflow through AI-powered automation. Luminai’s platform combines advanced AI and machine learning to automate complex, multi-step workflows across and within teams such as revenue cycle management, elimination of faxes, eligibility and intake, claims processing, and other mission-critical functions. Health systems using Luminai have achieved immediate and measurable results — accelerating processing times, improving accuracy, and removing bottlenecks that delay care and payment. By reducing the time clinicians and staff spend on repetitive tasks, Luminai not only drives significant cost savings and productivity gains but also enables healthcare teams to focus on higher-value work that improves patient experience and outcomes.

‍

Solara Health is tackling one of healthcare’s most urgent challenges, our clinician workforce crisis. Starting in behavioral health and expanding across clinical specialties, Solara’s AI-driven platform enables clinicians to focus on clinical care while providing support so they can practice at the top of their licenses. Solara’s solution ingests a health system’s own clinical protocols and evidence based guidelines and deploys clinical tools customized for each health system. Examples include recruiting assessments, provider training modules, patient facing CBT modules, voice enabled patient intake, among many others. As one of the first companies to leverage large language models for both clinical and operational challenges, Solara is increasing the joy of clinical practice while allowing providers to be more efficient with their patients. 

‍

The power in this framework is in showing how these layers interact. Personalization at the front door relies on a strong foundation of data. Insight extraction in the hallways enables smarter automation in the rooms. Automation in turn frees up capacity for deeper personalization.

‍

Founders who understand the house as a system — rather than four disconnected opportunities — position themselves to move from wedge to platform. They may enter through a single door, but the long-term vision is to build something that touches multiple layers of the house, creating defensibility and enduring value.

‍

Strategics, too, benefit from this lens: it clarifies where AI is likely to create near-term impact versus where long-term transformation will require re-architecting the entire house. It’s a framework for prioritizing investments, partnerships, and internal build-vs-buy decisions in an environment where everyone is racing to “do something in AI” but few are yet connecting the pieces.

‍

Across the house, we’ve also made several stealth investments in AI-native companies pursuing similarly bold visions including:

‍

• Expanding access to care for patients and unburdening clinicians and healthcare workers with Generative AI voice and chat assistants.
• Connecting community research sites with pharmaceutical companies, streamlining clinical trials from site discovery to completion.
• Accelerating drug discovery and development by scaling AI agents that work collaboratively with human researchers
• Reimagining the pharmacy experience for both health systems and patients.

‍

What Winning Looks Like — And What it Takes to Get There 

‍

AI is everywhere in healthcare right now — and yet meaningful adoption still feels just out of reach.

‍

Founders are under pressure to position themselves as “AI-first” but struggle to break through the noise of a crowded market where every pitch sounds the same. Meanwhile, healthcare leaders face the opposite challenge: relentless vendor outreach, lofty promises of transformation, and no clear signal on which solutions will truly move the needle.

‍

The last 18 months were defined by experimentation — every board wanted to “do something in AI,” leading to a wave of pilots across payers, providers, and pharma. That era is ending. The next phase is about execution: fewer experiments, deeper commitments, and measurable results.

‍

Three significant shifts stand out: 

‍

• Fewer, deeper partnerships. Rather than juggling dozens of point solutions, enterprises are starting to prioritize platforms that can scale across multiple workflows. For example, the majority (56%) of pharma companies said they plan to partner with select vendors that address multiple use cases. 

‍

• Hard ROI over soft learning. Early “exploratory” budgets are giving way to demands for quantifiable outcomes. Today 61% of payers and providers said establishing ROI is their top integration challenge, demonstrating the shift towards cost savings, revenue impact, clinician efficiency, etc. 

‍

• Multi-stakeholder sales by default. Purchasing decisions now require alignment across clinical, operational, and financial leaders — and increasingly through formal governance bodies with 80% of pharma companies and 73% of payers and providers with established AI governance committees. This introduces both complexity and clearer pathways for enterprise-wide adoption.

‍

Startups vs. Incumbents — and the New Build-vs-Buy Debate

‍

Startups today face a dual challenge: competing with both healthcare incumbents “AI-ifying” existing processes and enterprises accelerating internal builds, especially for well-defined operational use cases. The speed of model development makes in-house experimentation more viable than ever.

‍

Yet this shift also creates opportunity. Startups are structurally advantaged in reimagining the value chain itself — unconstrained by legacy workflows or institutional inertia. As one health system leader described, partnering with a startup is like “having dedicated compute power focused on our hardest problems” — high-velocity teams that innovate in quarters rather than years.

‍

This reframes the classic build vs. buy question into “could vs. should.” Healthcare organizations could build internally, but should they? For high-stakes problems that span multiple stakeholders or require deep AI expertise, partnering with external teams often proves faster, cheaper, and more scalable. Especially when those startups are architected for compliance, workflow integration, and enterprise readiness from day one.

‍

Winning companies in this environment share five traits:

‍

From wedge to platform. Winning companies don’t just find any wedge — they identify a use case urgent enough to drive adoption and defensible enough to create early traction. But they don’t stop there. In today’s AI environment, where technology and buyer expectations move in quarters rather than years, founders must define their platform vision far earlier and with far greater clarity. It’s no longer enough to solve one pain point; they must articulate how their technology will expand across adjacent workflows and stakeholder groups, and know their roadmap well enough to execute as new opportunities emerge.

‍

Relentless customer curiosity. Winning companies deeply understand every relevant audience for their solution — from frontline clinicians and researchers to the operational leaders and governance committees that ultimately decide if a product scales — and they pair that understanding with a commitment to being a great partner. This means anticipating adoption barriers, designing workflows that fit real-world needs, and building trust through consistent delivery.

‍

Workflow integration as defensibility. The most successful companies build products that integrate seamlessly into clinical, scientific, and operational processes — embedding insights where decisions are already made rather than asking users to log into yet another platform. This is where multi-stakeholder complexity becomes a moat: solutions that create value across roles (e.g., clinicians, researchers, business development, coding teams, governance committees) are harder to displace and scale more naturally across the enterprise. 

‍

“The question is not how do we improve a cumbersome process, it’s do we need to do the process at all?” — David Reese, CTO of Amgen

‍

Clockspeed in a fast-moving market. Technical moats erode quickly, but the ability to adapt decisively becomes a core advantage. Winning companies evolve their product roadmaps in lockstep with market shifts, integrating new capabilities while staying grounded in regulatory, workflow, and stakeholder realities. It’s not just about moving fast, it’s about moving smart and with purpose.

‍

Cross-disciplinary depth. The strongest teams combine world-class AI talent with deep healthcare expertise. Increasingly, we’re seeing repeat founders pairing with leading researchers — blending technical credibility with commercial intuition from day one.

‍

The winners will be those who can meet this moment — not just by building great technology, but by navigating healthcare’s complex buying dynamics, outpacing incumbents and internal builds, and earning the trust required to scale.

‍

A Call to Shape the Future

‍

At Define Ventures, we see ourselves as the connective tissue across the healthcare ecosystem — linking founders building at the frontier of AI with the incumbents that can scale their impact. 

‍

For founders, that means close partnership through the hardest stages of company-building: identifying urgent wedges, navigating enterprise sales, and evolving into platforms that transform care delivery. For strategics, it means a trusted collaborator to help cut through the noise, prioritize high-impact use cases, and forge partnerships that accelerate innovation. And for the industry at large, it means keeping our ear to the ground — continuously refreshing our thinking as the technology and market evolve.

‍

The convergence of systemic pressures, technological readiness, and commercial appetite makes this a once-in-a-generation moment. Healthcare will not be transformed overnight, but the foundations are being laid now. We believe the winners of this era will be those who harness AI not as hype, but as a tool for lasting impact. If you’re building an AI-native healthcare company with ambition beyond the wedge, or a strategic seeking to partner on transformative solutions, we’d love to hear from you. This next phase belongs to those willing to reimagine healthcare from the ground up.

‍

Greg Brockman reminded us at Define’s AI Summit that “there is no finish line” in generative AI . Instead, it is a spectrum of ongoing innovation. The same is true for healthcare. The real question isn’t whether AI will transform the industry; it’s who will actively shape that transformation or simply react to it.