Isaac Park, Co-Founder & CEO of Keebler Health, on why $16M is just the beginning of fixing the most expensive blind spot in healthcare

Recently, we closed $16 million in Series A funding, led by Flare Capital Partners with participation from Sands Capital and several other incredible investors.  We’re proud and grateful for the trust our investors continue to show in us, and while the dollars raised might be the headline, we believe the real story lies with what this will do for improving patient outcomes.

Our lead investor in this round, Flare Capital, has spent decades investing in companies that have successfully moved healthcare forward.  Sands Capital has an even older reputation, being active supporters of transformational growth at scale, particularly in the health and life sciences sectors.  Having both of these storied funds in our corner means something significant: that the transformative healthcare opportunity we see is real, the timing of what we’ve built is right, and the results we’ve produced are worth betting on. 

With these incredible backers behind us, we’re aiming to drive our vision of better patient outcomes; we are rapidly growing our team, expanding platform capacity, and forging deeper partnerships with value-based care enablers.  Most importantly, we’re proud to support providers motivated to leverage clinical narrative as the linchpin for preventive care.

Healthcare Data: a Poor Proxy for Health

When Andrew, Kevin, and I started Keebler Health, LLMs were just beginning to reveal themselves as more than a research curiosity.  And up to that point, genuinely useful healthcare AI tools had real value (reading lab values, parsing claims, and predicting structured results), though they were constrained in their usefulness.  Structured data (as good as our industry had gotten at using it) could only ever be a limited proxy for what was happening in a patient’s clinical journey.

The clearest story of a patient’s health never actually lives in a sterilized claims history; instead, it lives in clinician documentation – all of it, everywhere, together: the discharge summary, the offhand mention of a symptom buried on page 47 of a specialist visit, the telephone note to a medical assistant.  It is all of that narrative, used together, that clinicians heroically consume on a daily basis to build a mental model on how to care for their patients.  

And it was that found inspiration that was the most exciting insight from our early work innovating in this space.  We realized that clinicians never had the appropriate software tools to really amplify what they already were doing, building that same clinical mental model of a patient’s journey by reading the actual documentation.  That practice, at best, was a manual job, never achievable at scale.

This is all different now.

For the first time in history, humankind has the technology to read, metabolize, and inference over unstructured narrative at the speed and scale of the internet.  Gone are the days when determinism limited the ability for disparate systems to interact with one another electronically. Instead, orchestration over legacy products can actually happen over the primary protocol that almost all software infrastructure already uses: natural language.

And I know of no other vertical, certainly not one as large as the $6 trillion dollar U.S. healthcare market, built entirely on a bedrock foundation of unstructured narrative.  In other words, healthcare suddenly became the largest untapped opportunity for technology in what feels like a lightning strike moment in history.  And in our case, we’re harnessing that energy to empower our VBC partners to deliver better patient care.

How We’re Doing It

It’s not lost on us that almost every healthcare software company suddenly became an AI company overnight.  It makes sense; we would have done the same had we needed to take inherited architecture and figure out how to layer newer intelligence on top of it.  Our infancy, however, became our advantage.  With no existing infrastructure to have to work around, we were able to build specifically for our vision from the very first line of code.  And our architecture reflects the design of how clinical intelligence actually should work, not what a prior generation of software happened to support.

In practice, that means we process every data source together: structured and unstructured, tidy and messy, exact and ambiguous.  Just like a clinician, we can correlate the nuance across all of the data, nuance that used to get scrubbed out so deterministic systems could function.  An offhand mention, an implied diagnosis, or the buried symptom on page 47 – all left off of any calculus of the patient story.  We don’t do that.  The nuance is the point. 

We surface chronic disease evidence buried deep in the documentation, and we connect every finding to the note that defends it so there’s always a traceable, auditable trail. When CMS updates its models, we move with it, because a platform that requires manual reconfiguration every time the rules change isn’t one that care teams can actually build on.

For coding teams, this changes the nature of the work entirely.  What has historically taken weeks now takes minutes; and not because we automated a task, but because we fundamentally changed what the task is.

We also close the gaps that fragmentation quietly creates.  A condition documented in a specialist note that never made it into the primary care record.  A diagnosis implied by a medication list that no coder ever formally connected.  We find those threads and follow them, until what emerges is a coherent, longitudinal picture of a patient’s journey: the kind of picture that value-based care was always supposed to be built on, and rarely actually had.

Chronic Disease Detection is Just the Start

When clinical complexity is under-documented, the consequences aren’t abstract.  Clinicians get undercompensated and access narrows.  Care teams absorb the slack until they can’t, and who pays the ultimate price but patients?  In other words, poor risk adjustment isn’t a billing optimization problem.  When done in the spirit of how it was always intended to be done, bad risk adjustment is actually a clinical equity problem, and I think our industry has been too comfortable calling it a reimbursement issue to avoid the weight of it being a clinical quality issue.

Thankfully, we’re surrounded by a host of VBC enablers that see it the same way, including our administration.  CMS has stated its goal of moving all Medicare and Medicaid beneficiaries into value-based arrangements by 2030.  And the organizations that will thrive in that world aren’t the ones that just made their existing workflows incrementally better.  They’re the ones that deploy an operational layer that actually truly reflects the spirit of what Value-Based Care actually is: a deep coupling of financial and clinical outcomes, or in other words, deploying a technology that bridges clinical unstructured narrative truth with how VBC organization should actually operate.

And that’s what we’re building.

When you can read the full clinical narrative, you stop reacting and start anticipating.  A lapse in medication adherence becomes visible before it becomes a hospitalization.  A pattern across a population becomes legible before it becomes a crisis.  It is the kind of medicine that clinicians actually want to practice: proactive, informed, and actually responsive to what’s happening in front of them.

I’ve spent most of my career at the intersection of technology and healthcare, and the problem has always been the same one: the data has always been there, but the tools to read it at scale never did, until now.  Again, I know of no other industry where that shift arrives with more significance or with more urgency, and I feel incredibly privileged to be a participant in this moment.

To our investors, our customers, and our team: you already know what this means. The best chapters are still ahead.

If you’re a health system serious about risk adjustment, we’d love to talk.  And similarly, if you want to help with this work, please don’t hesitate to reach out!  We’re hiring!

— Isaac Park, Co-Founder & CEO, Keebler Health