Charta Health raises $55M to expand AI coding for insurers

Charta had already announced a $22 million Series A led by Bain Capital Ventures on July 22, 2025, describing software that augments or replaces manual chart reviews. The new round will fund further AI development for provider teams and expansion into additional specialties, as well as the insurer suite.

The company says in its launch press release that revenue grew tenfold over the preceding 12 months across health plans and provider organizations. It does not disclose the revenue amounts behind that growth multiple. The payer launch therefore builds on existing insurer work, rather than marking Charta's first business with health plans.

One engine, different review workflows

Charta's pitch rests on large language models reading unstructured clinical records, including free-text notes, scanned orders and lab results. For providers, Charta says its software reviews every chart before claim submission against coding guidelines, payer rules and the customer's clinical protocols. Findings appear inside the electronic health record for human approval. Insight describes simple corrections as automatic and more nuanced issues as routed to clinicians.

This requires more than reading clinical prose. Insight says Charta configures its engine to each customer's rules across specialties, states and payers. The investor also describes a fragmented electronic-record market and says customers reported deployments in weeks and rule changes within days. Those are investor-reported customer experiences, not a published implementation benchmark.

Charta names American Family Care, Grow Therapy and Pediatric Associates Family of Companies as customers. The financing report also describes an agreement earlier this year to offer Charta through the athenahealth Marketplace, an app and integration ecosystem for providers using athenahealth's electronic records and billing software.

Charta for Payers initially targets Medicare Advantage and Affordable Care Act Marketplace plans. Charta says in its product announcement that it designed the suite with one of the country's largest health insurers, which it does not identify. Its six products cover:

• Retrospective diagnosis coding from historical charts.
• An automated evidence check on first-pass coding.
• A workspace for coders to review and finalize findings.
• Pre-visit information about conditions to address at the point of care.
• Validation of submitted diagnoses and workflows for risk-adjustment audits.
• Detection of gaps in quality-of-care measures from chart data.

The company says the engine identifies both documented conditions omitted from coding and coded diagnoses unsupported by the chart. That two-way check is the substance of its accuracy pitch. But historical reviews and audits of submitted diagnoses also show why the suite should not be described simply as pre-billing software.

Charta enters an existing automation market. The financing report identifies Arintra's AI revenue-assurance platform, Candid Health's provider revenue-cycle platform and CodaMetrix's AI medical coding business. Their functions are not identical. On the insurer side, Medicare's payment advisers had already documented plans and vendors using AI to select charts likely to generate additional revenue. Charta's proposition is a common review engine for both sides; automated coding and insurer chart review already exist.

MedPAC estimated that chart reviews alone accounted for about $24 billion in 2023 Medicare Advantage payments, using 2022 encounter records. This does not measure fraud or Charta’s performance. Graphic: Magica. Source: MedPAC March 2025 Medicare Advantage report.

Better coding can raise government payments

Medicare Advantage lets Medicare beneficiaries receive benefits through private plans instead of traditional fee-for-service Medicare. Medicare pays those plans a rate per enrollee adjusted for expected medical costs, using risk scores based partly on diagnoses. Recording an additional eligible diagnosis can therefore increase a plan's payment without changing the care delivered.

The Medicare Payment Advisory Commission, Congress's adviser on Medicare payment policy, offers a direct measure of the stakes. In its March 2025 report, it estimated that diagnoses found only through chart reviews accounted for about $24 billion in 2023 payments to Medicare Advantage plans. Its analysis used 2022 encounter records to identify diagnoses supported by chart reviews but not another physician or hospital encounter record during that year, then estimated the associated payments.

This is not an estimate of fraudulent payments or Charta's addressable savings. A chart review can identify a valid diagnosis omitted from the usual reporting process. MedPAC says differences in coding between Medicare Advantage and traditional Medicare reflect more comprehensive documentation, fraudulent submissions and other factors whose effects cannot always be separated.

The report also projected that higher coding intensity would increase 2025 Medicare Advantage payments by $40 billion relative to what spending would have been for comparable beneficiaries in fee-for-service Medicare, after Medicare's coding adjustment. That projection excludes beneficiaries with end-stage renal disease and is a system-wide estimate, not a Charta performance measure. Taxpayers and Medicare beneficiaries finance those higher payments.

The distinction matters for assessing the product: finding a supported but omitted diagnosis can raise a plan's payment; flagging an unsupported diagnosis can lead to its removal and lower payment. Both can improve accuracy. Neither alone demonstrates lower administrative costs or fewer disputes.

What buyers still need to establish

There is evidence of a working provider business, but the performance disclosures have limits. In its Series A announcement, Charta described pediatric home-health provider KidsCare Home Health as reducing clinical-management workflow duties by at least 25% across 19 clinical managers and saving an estimated 10,000 hours annually. That is a named customer example supplied by Charta, not an independently evaluated result or a measured saving across insurers and providers.

Charta now claims consistent double-digit reductions in provider denial rates. Its financing announcements do not supply the sample sizes, measurement periods or methods needed to assess how broadly that claim applies. Nor do they disclose product prices or the integration and human-review costs needed to calculate a buyer's net savings.

The next decision for health plans and providers is whether the expanded suite delivers enough verified benefit to justify adoption. Results would need to distinguish supported diagnoses added from unsupported diagnoses removed, disclose human review and implementation costs, and track subsequent denials and disputes. Charta has capital to expand an existing business; evidence that its shared engine produces a shared reduction in billing friction remains the test.