The Moat Moved While Everyone Was Looking at the Model

Some important AI products will be invisible to patients

Revenue-cycle management, coding, prior authorization, documentation, capacity planning and diagnostic workflow rarely make for the most exciting product demo. They sit inside expensive, recurring work.

I increasingly ask who is accountable for the result. A hospital CFO needs to see collections or cost improve. A physician needs to know whether the output changes care. An administrator needs an implementation the organization can actually support.

The buyer, the user and the person carrying the risk may be three different people. Understanding all three is part of understanding the business.

Regulatory discipline belongs in the same discussion. Data provenance, validation, auditability and change management affect whether an institution can trust a product enough to use it repeatedly. They shape product design and adoption from the beginning.

The six questions in our underwriting

Our work keeps returning to a small set of questions:

• Does normal use create a proprietary advantage? I want to see what improves, who owns it and why a competitor cannot reproduce it easily.
• Does the product change a consequential decision? Screening, diagnosis, treatment, monitoring and reimbursement each create different demands for evidence.
• Does it live inside a recurring workflow? A customer trying a product once tells us much less than a team depending on it every week.
• Can technical performance survive the journey to adoption? Validation, implementation, reimbursement and distribution need to be part of the plan.
• Are the data rights, IP and regulatory claims defensible? Access to data and durable rights to use it are separate questions.
• Who pays today, and who may eventually need this capability? We cannot predict an acquisition, but we can study strategic dependence.

That last question is worth sitting with. Which organizations become stronger if they own the technology? Which become more vulnerable if a competitor does? The answers can reveal something the market-size slide misses.

Why I want disagreement around the table

A scientist may love a result while a physician questions whether it changes care. An operator may see the clinical value and still struggle to picture implementation. A reimbursement expert may find the economic problem everyone else missed.

Those are the conversations I value most. They are also why we are building Succedia’s work around people with experience across clinical science, data and product, healthcare operations, commercialization, reimbursement, IP and venture underwriting.

An impressive adviser list does little by itself. I want someone in the room willing to explain precisely where an assumption breaks. Sometimes that saves us from a poor investment. Sometimes it reveals the part of a company we had underestimated.

I feel fortunate to keep learning from people who will challenge me that way. Legacy grew out of wanting more of those conversations to happen, and more of what we learn to reach the wider community.

What we are comfortable missing

We will see companies succeed outside our focus. That is part of having one.

We remain cautious about products whose main advantage is access to widely available models, biomarkers with no clear decision attached, dashboards that never become part of the workflow, and excellent science without a credible route to adoption and payment.

The next company we want to understand early may sit in a pathology lab, behind an imaging workflow or inside software few people outside a hospital ever see. It will do something difficult repeatedly, gather evidence that it works and become a capability other people depend on.

That is where I’m spending more time going into Q4.