US healthcare · Denial appeals · Billing companies and specialty practices
Nobody appeals the small denials. That is where the money is.
Between half and two-thirds of denied medical claims are never reworked — not because they are unwinnable, but because working one costs more in staff time than the claim is worth. AppealIQ is being built to make appealing every denial economical: read the denial, work out why it happened, draft the appeal with the payer's own policy cited, file it, and follow it to a decision. A person approves every letter before it leaves.
$52,800
written off without anyone reading the denial
- Appealed and overturned15%
- Appealed and upheld21%
- Never touched64%
Industry averages applied to a representative practice, not our results. We have not run a single appeal through a product yet — see where this actually is.
The problem
Payers deny with software. Practices reply with spreadsheets.
A denial is not a verdict, it is an opening position. Appeals win often — 40 to 70 per cent overall, and 80 to 90 per cent on the purely technical ones. But a small practice triages by dollar value, because each appeal costs real staff hours, so everything under an internal threshold is written off unread. The threshold is the product opportunity: collapse the cost of one appeal and it becomes rational to fight all of them.
$262B
of claims initially denied every year in the US.
Change Healthcare Denials Index
118
dollars, at the top end, to work a single denial by hand — $25 at the bottom.
MGMA
65%
of denied claims are never reworked at all, up to this share.
MGMA · 50–65% range
Figures from public 2024–2026 industry sources: Change Healthcare, Kodiak Solutions, MGMA, the AMA physician survey (Feb 2025) and Bruch et al., JAMA Internal Medicine (Apr 2026). They are directional, and they describe the market — not any outcome we have produced.
What AppealIQ does
Three things, end to end.
The point is not a worklist. Software that tells a billing manager which denials to fight has moved the problem, not solved it — the expensive part was always the fighting.
Reads the denial
Takes the remittance file, the claim and the clinical documentation, and works out what actually went wrong — a missing authorisation, a coding mismatch, a policy the payer applied incorrectly, or a denial that is simply correct.
Writes the appeal
Drafts a letter aimed at that payer, in that state, for that specialty, citing the policy language it is arguing against and assembling the evidence packet that has to travel with it. Then it puts the whole thing in front of a human.
Files it and chases it
Submits by whatever route the payer actually accepts, tracks the claim to a decision, escalates to the next appeal level when one is available, and reports what came back in dollars rather than in tickets closed.
The four commitments
What we are accountable to.
This product touches protected health information and speaks to insurers on a practice's behalf. These are the constraints it is being built inside, not features to be traded away when a demo needs to look faster.
- A human approves every appeal Every letter goes through a review queue before it is filed. Auto-approval is earned, category by category, only after accuracy has been measured against real outcomes — and it is off on day one.
- Citations are checked, not trusted A fabricated policy reference sent to a payer is worse than no appeal at all: it costs the practice its credibility, not ours. Every citation is verified against the source before a letter can be approved.
- It will tell you not to appeal Some denials are correct, and a tool that appeals everything wastes the staff time it claims to save. "Correct this and resubmit" and "write this off" are valid answers the agent is expected to give.
- HIPAA posture from day one Business-associate agreements, encryption, least-privilege access and an access log from the first line of code, with a per-appeal audit trail as a product feature rather than a compliance chore. SOC 2 follows on a stated timeline, not a vague one.
Deliberately not shown
We are not putting the hard part on a website.
Which argument, which evidence, and which phrasing overturns which denial code — at which payer, in which state, for which specialty — is the entire product. That data does not exist publicly, cannot be bought, and is the only thing here a foundation model will never have. Design partners see the mechanics under NDA. Everyone else gets the paragraph above.
Withheld until launch
- Payer policy library
- Appeal-outcome dataset
- Verification layer
- Submission mechanics
Nothing is hidden behind those bars — there is no text underneath them to reveal.
Where this actually is
Validation phase. No product yet, and we are saying so.
Every other page on this site links to something you can open. This one does not, because AppealIQ has not been built. We are doing the unglamorous part first — proving the money is recoverable in someone else's real data before writing a line of the thing that recovers it.
Before we write code
- Interview 15 to 20 billing-company owners and quantify, in their words, how much they write off unappealed and why.
- Get de-identified remittance files and hand-work 20 appeals ourselves, to measure an overturn rate we have actually produced.
- Confirm, payer by payer, how an appeal has to be submitted in the target specialty and state. Some of it is still fax.
- Price-test against the recovered dollars found in that audit, rather than against a competitor's page.
- Bring in a revenue-cycle advisor who has run a billing company, so the policy library is credible on day one.
Pre-committed pivot trigger. If the appeals we hand-work overturn at under 40 per cent in the chosen specialty, we change the specialty or the payer mix before building anything. Writing that down in public is the point of writing it down.
Who it is for
Small billing companies first, then the practices.
Billing companies
Ten to two hundred provider-clients, paid a percentage of collections, so a denial written off is margin lost twice. One conversation reaches dozens of practices, which is the only distribution that works in a market of two hundred thousand of them.
Independent specialty practices
Five to fifty providers in the specialties payers deny hardest — orthopaedics, behavioural health, GI, cardiology, dermatology, DME — reached through specialty societies rather than cold outbound.
Hospitals and health systems
Already served, at length, by enterprise revenue-cycle vendors. We are building for the long tail those platforms price out and ignore, and going up-market later is a different company.
Design partners
Bring us your denials. We will tell you what is recoverable.
We are looking for two or three billing companies or specialty practices willing to hand over de-identified remittance data for a free denial audit: what was denied, what was abandoned, and what we think is winnable — in your own numbers, before you owe us anything.