The Case for a Neutral Healthcare Settlement Exchange That Turns Payer-Provider Contracts Into Executable Code, Guarantees Payment, and Makes Most Denials Structurally Impossible
Video Preview
🎧 Part I Podcast free on Spotify.
🎧 Part II Podcast episode for paid subscribers only. Also available on Spotify.
To listen to paid episodes in Apple or Spotify, link your Substack subscription via the show settings on those platforms (instructions inside the Substack app under Subscriptions → Podcast).
Abstract
Thesis: the next enormous healthcare infrastructure company won’t win the denial war with better AI. It will retire the war by giving payers and providers one shared, executable version of their contract and settling against it.
The core defect: today each side compiles the same signed contract into different software. Two interpretations, one agreement, endless variance. That gap is where denials, appeals, underpayments, and days in AR live.
The product being sold is finality, not workflow. Both parties agree in advance that the exchange’s output is the presumptively correct payment.
Money comes from five places: a per-claim settlement fee, a share of verified admin savings, a payment-guarantee and financing spread, mutual performance bonds, and paid micro-arbitration for the genuinely uncomputable cases.
Why now: recent academic and industry work argues existing standards (FHIR, CQL, FHIRPath, X12, plus domain rules) can represent nearly all real contract terms, and recent surveys show many payers still run contracting on spreadsheets, with only a minority using any contract analytics or AI review.
Beachhead: self-funded employers and direct contracts, then state Medicaid programs, then delegated-risk arrangements. Not national insurers on day one.
Moat: a canonical contract ontology, versioned reimbursement modules, bilaterally validated logic, historical test suites, exception-resolution data, network effects, and the capital to guarantee payment. The model that extracts contracts is not the moat and will commoditize.
Structure: has to be neutral. Public-benefit corp, split payer-provider governance, segregated reserves, audited rules. The Change Healthcare outage is the cautionary tale for concentration.
Table of Contents
1. The arms race nobody actually wins
2. Two copies of the same contract, and why that’s the whole problem
3. What the exchange really sells is finality
4. How the thing makes money
5. Why this is buildable now and wasn’t a few years ago
6. Why the incumbents didn’t already do it
7. Where to plant the flag first
8. The payer question, which is the only question that matters
9. The moat, which is not the language model
10. Neutrality as product, and the ghost of Change Healthcare
11. What it becomes if it works
The arms race nobody actually wins
Walk into any revenue cycle shop right now and the pitch is the same. Buy our model, catch denials before they happen, auto-appeal the ones that slip through, recover the underpayments, shave a few days off AR. Walk into the payer side and the pitch rhymes. Buy our model, tighten the edits, flag the outliers, route the suspicious stuff to review, hold the line on medical policy. Everybody is buying compute. Everybody is very proud of the compute. And the net effect on the actual system is close to zero, because both sides are pointing that compute at each other across a gap that shouldn’t exist in the first place.
That is the part worth sitting with. Providers are building AI to predict, prevent, and appeal denials. Payers are building AI to edit, review, and investigate claims. Two armies, better weapons every quarter, and the battlefield itself never changes. It is a genuinely funny situation if you squint, because the thing they are fighting over is not some deep clinical mystery. Most of the time it is a math dispute about a contract both parties already signed. Eight hundred dollars. Nine months of letters, calls, portal logins, faxes that somehow still exist, and a couple of full-time humans on each side whose entire job is to argue about eight hundred dollars that the contract already resolved if anyone bothered to read it the same way.
The interesting move is not to build a slightly better soldier. It is to notice that the war is optional. If a payer and a provider could agree, once, on a single machine-readable version of what they signed, and agree that the output of running a claim through that version is the correct answer, roughly the entire denial-and-appeal industrial complex loses its reason to exist. Not because anybody got nicer. Because the ambiguity that funds the fight got compiled away.
That is the whole idea. Everything else is plumbing, incentives, and go-to-market, which is to say everything else is the hard part. But the concept is clean, and clean concepts in healthcare are rare enough that it’s worth taking this one seriously instead of reflexively filing it under “yet another contract platform.”
