TrueClaim reads every medical claim before your company pays it

If your company self-funds its health plan, and once a business gets past a few hundred employees it almost certainly does, here is an uncomfortable fact about how that money moves.

Every month, your health plan pays out a stack of medical claims. Doctor visits, lab work, imaging, surgeries, prescriptions. Each one is a bill that a hospital or a pharmacy sent in, and each one is paid with your company's own money, because a self-funded employer is the one actually on the hook for the cost of care. Its members carry an insurance card, but the insurer name on that card is usually just an administrator processing paperwork. The dollars are the employer's.

And almost none of those bills get read. They get processed. A third-party administrator receives the claim, runs it against the plan rules, and cuts the payment. If the hospital billed for two of something that happened once, or coded a routine procedure as a complicated one, or charged for a service the plan doesn't even cover, the payment usually goes out anyway. The employer finds out at the end of the year, when the total comes in higher than last year, and nobody can say exactly why.

That gap is what TrueClaim is built to close.

TrueClaim is a transparent, AI-enabled third-party administrator for self-insured companies. It sits in the payment flow between healthcare providers and the employers who fund their own plans, and its software adjudicates and continuously reviews 100% of medical claims before the money leaves. The company's Y Combinator launch put the pitch more bluntly: the world's first transparent, AI-enabled healthcare TPA that doesn't suck. The claim underneath the joke is a serious one. TrueClaim says it can save an employer at least 7% of its healthcare spend without changing anyone's benefits, and as much as 20 to 30% when the employer acts on its benefits-design suggestions.

The part of your benefits budget nobody is watching

To see why this is a real opening, you have to understand how much of American healthcare actually runs on the self-funded model.

When people picture employer insurance, they picture a company buying a policy from a big insurer and paying a premium. That is how small businesses work. But most large employers don't buy insurance at all. They self-fund: they pay for their employees' claims directly, out of company cash, and hire an administrator to run the plan. The insurer logo on the card belongs to whichever company won the administration contract. It is a services arrangement, not a risk-transfer one. The employer keeps the risk and the bill.

Self-funding can be a good deal for a company, because it gets to keep whatever it doesn't spend instead of handing an insurer a margin. But it also means the employer owns every mistake in the claims stream. And the claims stream has a lot of mistakes. Industry studies of employer plans routinely put claim error rates somewhere between 2 and 6%. On a health budget that runs into the millions for even a mid-sized company, a few percent of errors is real money, paid out quietly, month after month, with no one on the employer's side reading the line items.

The reason no one reads them is structural. A benefits team of two or three people cannot manually audit tens of thousands of medical claims a year. So they trust the administrator to get it right, and they look at the total once a year when renewal comes around. The work that would catch the errors is exactly the work that never happens.

Why the incumbent administrators don't fix it

The natural question is why the existing administrators haven't solved this already. The answer is that many of them are not especially motivated to.

The traditional administrator business has quietly grown a set of incentives that don't line up with the employer's. Some administrators take hidden fees out of the claims flow. Some engage in spread pricing, where they pay a provider one amount and bill the employer a higher one, keeping the difference. And in one of the more perverse arrangements the industry has produced, some administrators charge the employer a percentage fee on the "savings" they recover, which means they can earn more by letting an improper payment go out first and then collecting a cut when they claw it back. Legal cases against several large administrators have described versions of all of this.

None of that is illegal on its face, and not every administrator does it. But it explains the inertia. If your revenue depends partly on the friction and opacity of the claims process, you have little reason to make that process transparent or to catch every error the first time. The employer, meanwhile, rarely has the visibility to notice. This is the incumbent problem TrueClaim is positioning against, and it is a genuine one. The point isn't that administrators are villains. It's that the model was never designed to be inspected, so it isn't.

TrueClaim's argument is that the whole thing can be rebuilt around the opposite assumption: that the employer should be able to see every member, every claim, and every dollar, and that the administrator should make money by lowering the bill rather than by working the seams of it.

What the software actually does

TrueClaim's core is an engine that adjudicates and continuously reviews all of a plan's medical spending, not a sample of it.

Because the review is software rather than a team of auditors, it can run against 100% of claims and keep running. TrueClaim describes its system as AI agents that work around the clock, checking claims for billing errors, adjudication mistakes, and gaps in care. The engine draws on three kinds of information that have only recently become usable together: the medical claims themselves, the clinical documentation behind them, and the price transparency data that hospitals and insurers are now required to publish. Put those alongside each other and a lot of problems that used to be invisible become checkable. Was this procedure billed twice? Is the charge in line with the published rate for this provider? Was a cheaper, equivalent drug available? Did a member fall through a care gap that will turn into a bigger claim later?

TrueClaim groups the savings it finds into a few buckets. There are billing errors, the duplicate charges and miscodes that should never have been paid. There is pharmacy, where the same medication can carry wildly different prices. And there is care navigation, steering members toward care that is both better for them and less expensive for the plan. The first bucket is about not overpaying for what already happened. The others are about shaping what happens next.

On top of the engine sits a real-time dashboard that lets a benefits team actually watch the plan, member by member and claim by claim, instead of waiting for a year-end report. That visibility is part of the product, not a reporting afterthought. For a benefits leader who has spent years flying blind between renewals, being able to see the plan as it runs is a change in kind.

Getting started, in TrueClaim's telling, is deliberately low-friction: an employer introduces TrueClaim to its existing administrator or ASO, and TrueClaim plugs into the payment flow from there. The savings claim, at least 7% with no change to benefits, is meant to be the easy part of the conversation, because nobody has to give anything up to get it.

Transparency as the actual product

It would be easy to read TrueClaim as a cost-cutting tool with some machine learning attached. The more interesting way to read it is as a bet that transparency itself is the product.

The word "transparent" sits in the company's own description of what it is, and it is doing real work there. In a business where a good part of the incumbent margin comes from what the customer can't see, choosing to show the customer everything is a competitive stance, not a slogan. If TrueClaim shows an employer every dollar and makes its own money by lowering the total rather than by hiding fees inside it, then its interests and the employer's point the same direction. The administrator wins when the bill goes down. That is close to the reverse of how the spread-pricing, savings-fee version of the business works.

This alignment is also why an incumbent can't easily copy the approach. A cleanly transparent, error-catching administrator would cannibalize the exact revenue streams that make the traditional model profitable. TrueClaim carries none of that baggage. It was built from scratch around the assumption that the employer gets to see the whole thing.

Founders who have lived inside the problem

The team is a large part of why the pitch is credible, because this is a problem you mostly learn by living inside it.

Barbora Howell, the CEO, joined Hinge Health as its 30th employee and built and ran the unglamorous machinery of a fast-growing health company: HR, benefits, and customer billing, through a period when Hinge scaled past 700 people and hundreds of customers. She later ran clinical operations at Pine Park Health, and she holds an MBA from Stanford and an economics degree from Wharton. The relevant part isn't the pedigree. It's that she has personally sat on the employer side of benefits and billing, which is exactly where TrueClaim's customer sits.

Bobby Bayer, the CTO, spent seven years at the health-tech company pMD, where he built and grew its revenue cycle management software and services as an engineering manager. Revenue cycle management is the discipline of getting medical billing right at scale, which is the precise technical muscle TrueClaim's adjudication engine needs. The team also includes a physician-executive, which matters for the care-navigation side of the work, where clinical judgment and cost can't be separated.

That combination, someone who has run employer benefits operations paired with someone who has built billing systems for a living, is a hard one to assemble, and it maps almost exactly onto the two halves of the problem. TrueClaim raised about $4.1 million to go after it, including a $2.1 million pre-seed in 2024, and came through Y Combinator's Winter 2024 batch.

Why this is buildable now

A company like this could not have worked a few years ago, and the reason is data.

The AI part is real, but the quieter enabler is that the raw material to check a claim finally exists in the open. Hospital and insurer price transparency rules have forced a large amount of pricing into public view. Claims data and clinical documentation are more accessible in structured form than they used to be. Once all three sit in the same place, software can do what a human auditor never could at scale: look at every claim, compare it to what the care should have cost, and flag the ones that don't add up. The models are good enough to read the documentation and reason about it. The data is finally there to reason against.

TrueClaim's stated ambition is not small. The founders talk about saving 1.5% of US GDP while improving the quality of care, and they frame the opportunity as the roughly two million self-insured companies with fewer than a thousand employees that struggle to keep their healthcare costs in check. Whether or not the GDP figure ever lands, the shape of the bet is sound. There is an enormous amount of money moving through claims that no one is checking, and for the first time the tools to check it exist.

What an administrator becomes

For decades, the third-party administrator was a black box that employers paid to make the paperwork go away. You handed over the claims, you got back a total, and you hoped it was roughly right. The whole value of the arrangement was that you didn't have to look.

TrueClaim is betting that the value now runs the other way, that the administrator worth paying is the one that lets you look at everything and makes its living by lowering the bill. If that bet is right, the companies that win this category won't be the ones with the biggest networks or the slickest cards. They'll be the ones that read every claim before the money goes out, and can show you exactly what they found.