The digital health graveyard is filled with "groundbreaking" startups that had brilliant code but
zero clinical utility. In the high-stakes world of oncology, we see it constantly: a venture-backed
team develops a sophisticated algorithm in a vacuum, only to hit a brick wall when they try to
deploy it in a real-world hospital. They fall into the "pilot purgatory" trap, unable to scale because
they ignored the friction of the front lines.
Bridging this gap requires a fundamental shift in leadership- moving away from the visionary
technologist toward what Dr. Kamal Jethwani calls the "Concierge CEO." As the VP of Digital
Ventures at Moffitt Cancer Center and CEO of AccelerOnc Studio, Jethwani has spent two
decades proving that the difference between a failed experiment and a scalable company like
Decimal Health isn't the complexity of the AI- it’s the depth of the operational empathy.
The first hard-won lesson for any digital health founder is that technical novelty is a commodity;
operational implementation is the real moat. Most founders enter the oncology space obsessed
with their "black box" logic, yet they are blind to the "ground level" clinical environment. They
underestimate the sheer weight of technical debt and the "interoperability tax" that comes with
trying to force a new tool into a legacy hospital system.
Success is defined by how a tool survives the chaos of a busy oncology ward, not how it
performs on a clean dataset. To build a viable business, the founder’s mindset must shift from
"What can my code do?" to "How does this service solve a human bottleneck?"
"Operational implementation matters more than most founders realize."
To avoid building a product that nobody can use, Jethwani advocates for the "Concierge MVP."
This is a high-touch, human-centric phase where the product is essentially a service
masquerading as a platform. Before a single line of automated, AI-native code is scaled, the
"Concierge" (the startup team) manually performs the tasks to ensure the value proposition is
real.
This phase is critical for incubating the next generation of oncology companies. It allows
founders to:
Identify hidden friction points in data entry and retrieval.
Observe exactly where a clinician’s trust in the system breaks down.
Ensure the product provides immediate relief to a specific clinical burden before locking in the architecture.
We are entering a new era where oncology expertise is being productized through AI-native
companies built from the ground up. Institutions like Moffitt Cancer Center are no longer just
layering AI onto old systems; they are fostering startups that avoid the "interoperability tax" by
designing their entire core architecture around data liquidity and machine learning.
A non-negotiable pillar of this transition is Explainable AI. In oncology, transparency isn't a
feature, it’s a prerequisite for survival. The Concierge phase teaches founders that clinicians
don't need a "magic box"; they need a tool that offers:
Logical Transparency: Why did the AI suggest this specific clinical pathway?
Verifiable Data Sources: Which specific patient records or trial data informed the recommendation?
Actionable Confidence: Clear metrics that allow a physician to weigh the AI’s output against their own expertise
The Clinician’s Experience: The Workflow-First Mandate
For the clinician on the hospital floor, the greatest innovation in the world is a failure if it adds
five minutes to their rounding time. Digital health adoption is a zero-sum game of time and
cognitive load. The "hard-won lessons" of clinical adoption prove that there is a massive chasm
between a "good tool" and a "good workflow fit."
To achieve frictionless adoption, a product must meet the following criteria:
The Good Tool: Solves a technical problem (e.g., "It identifies a mutation").
The Good Workflow Fit: Solves a human problem (e.g., "It identifies the mutation and automatically flags the three most relevant clinical trials available at this facility, pre-filling the referral form").
Adoption succeeds when the technology is invisible and the utility is undeniable.
The transition from general digital health to specialized, AI-native oncology is well underway. Companies like Decimal.health represent this shift: moving away from "innovation forinnovation's sake" and toward a model where clinical expertise is baked into the code from day one. By embracing the "Concierge" model, founders can validate their impact at the ground level before scaling their technology.
The future of precision medicine isn't just about better algorithms; it’s about better integration. As clinical AI continues to evolve, the winners won't be the ones with the most complex code, but the leaders who master the art of the clinical workflow.
Are you building a tool that looks good in a pitch deck, or are you building a workflow that empowers a clinician at 3:00 PM on a Friday?