Candid Health: conquering the paperwork.

Somewhere in a multi-site medical group, a claim moves from a clinician's note to a payer's adjudication engine without anyone touching it. It is coded, validated, submitted, and paid. No phone calls. No follow-up. No spreadsheet. The billing manager finds out about it the way she finds out about most things now - in a dashboard, after the fact, while drinking coffee. This is what Candid Health is selling. It is also, depending on who you ask, either the dullest revolution in healthcare or the most important one.

The company sits at the intersection of medicine and money - an intersection where, until recently, every transaction seemed to involve a fax machine. Candid Health does not own a fax machine. It owns a rules engine, a set of pre-built integrations to electronic health records, and a thesis: that the United States' $280 billion annual spend on revenue cycle management is mostly a data engineering problem dressed up as a workflow problem.

Healthcare providers in the U.S. spend more than $100 billion every year just trying to get paid. McKinsey has estimated that streamlining claims submission could shave up to 18% off administrative costs - the kind of number that sounds bureaucratic until you remember it is bigger than the GDP of several countries. The reason it has not happened is not mystery. Payer rules change constantly. Codes are arcane. Denials cascade. Most billing software was built in an era when "interoperable" meant a CD-ROM and a prayer.

Nick Perry watched this up close. As Palantir's commercial lead for healthcare from 2012 to 2017, he sat with hospital systems and insurers and saw how much of the industry's "operations" was really just translation - clinical concepts into financial codes, financial codes into payer formats, payer formats into appeals. The work was tedious, expensive, and weirdly bespoke. It also produced extraordinary amounts of structured data that nobody seemed to be doing anything with.

In 2019, Perry left Palantir to start Candid with Doug Proctor and Adam Reis. The bet was unfashionable. Healthcare startups in that era were busy building shiny direct-to-consumer brands, telehealth pipelines, or AI scribes. Candid picked the unsexy middle of the stack - the part of healthcare nobody puts on their pitch deck. Y Combinator backed them. First Round, BoxGroup, 8VC, and eventually Oak HC/FT followed.

It is a peculiarly contrarian thesis. Most billing software vendors sell faster ways to do the same broken process. Candid's approach is closer to a rewrite: reverse-engineer the rules of every major payer, encode them in software, run claims through that engine before submission, and measure obsessively. The metric they care about most is "touchless claim rate" - the share of claims that go from clinical event to paid without a human in the loop. Most billing shops do not measure this. Candid centers its entire dashboard on it.

Strip away the marketing language and Candid is four things stitched together. There is a custom rules engine that knows what each payer will accept and reject. There are pre-built integrations into the EHRs providers actually use, so data flows in without anyone copy-pasting between systems. There is an analytics layer, more pleasant than most, showing net collection rate and cost-to-collect in real time. And there is a claims data export API for customers who want the granular numbers in their own warehouse - a small detail, but the kind of small detail finance teams notice.

The platform handles eligibility verification, coding assistance, claim submission, denial management, appeals, and payment posting. In Candid's telling, it is end-to-end. In practice it is more like a confident copilot that flags errors before submission and quietly fixes the ones it can fix on its own. The customer's billing team shrinks, or repurposes itself toward edge cases. The denial queue gets shorter. Cash arrives sooner.

It is one thing to claim a touchless claim rate. It is another to grow revenue 250% year over year, which is what Candid did in 2024 - the kind of growth investors will write checks for even in a market that has otherwise gone quiet on healthtech. Two rounds in six months is its own form of evidence. So is a roster of customers like Nourish, Allara, and Talkiatry, three of the more disciplined operators in modern digital health.

Candid puts it in three words: simplify medical billing. Behind those words is a more interesting argument. American healthcare does not need more clinicians fighting their software. It needs the software to stop being something clinicians have to fight. Every hour a billing team spends chasing a denial is an hour that could have gone to actual care, or to building the next product line, or to going home. The cost-to-collect ratio is not just a finance metric - it is a measure of how much of a hospital's day gets eaten by administrative drag.

The mission also explains the brand. In August 2025, Candid relaunched its identity around a Venn diagram - one circle for finance, one for healthcare, the overlap being the thing they actually sell. It is a quietly cheeky design choice. Most healthtech brands try to look like trust. Candid's new mark looks like a setup for a math joke. That fits.

The next few years of American healthcare will be defined by margin pressure. Provider groups are consolidating. Payers are tightening. Reimbursement is unkind. The systems that survive will be the ones that figured out how to get paid correctly the first time, every time, at low marginal cost. That is the bet Candid Health is making. If the bet works, the most boring software in the building becomes the most important.

Back to the scene. The claim has been paid. The billing manager has finished her coffee. Somewhere, an algorithm Candid wrote in 2022 just earned its salary for the year. Nobody noticed. That, in a sentence, is the point.