Ventures
Why Healthworx Bet on Defensible AI — And What It Means for Risk Adjustment

The investment landscape for healthcare AI has gotten more selective. Capital is moving toward platforms that can prove their value under scrutiny – not just in a demo, but in a federal audit. Healthworx's strategic investment in RAAPID, announced as part of the company's Series A extension, is a clear expression of that shift.
From code capture to defensible coding
For most of the last decade, risk adjustment technology optimized in large part for volume – more specifically – how many HCC codes could a platform surface from a chart? Accuracy was simply a marketing strategy. The question of whether a code could survive a RADV audit or a DOJ subpoena was rarely asked.
That calculus is changing fast. CMS has intensified its RADV audit schedule. The DOJ is pursuing False Claims Act cases against health plans whose retrospective programs added codes without removing unsupported ones. The OIG has published findings showing significant error rates at major plans. In this environment, a code that can't be defended isn't revenue, it's liability.
RAAPID was built around a different question from the start: not "can we find the code?" but "can we prove it?" Every diagnosis the platform surfaces is tied to a real clinical encounter, documented by a credentialed provider, and validated against MEAT criteria — Monitoring, Evaluation, Assessment, and Treatment. A chart note that mentions a condition with no supporting clinical activity gets flagged and rejected rather than passed along to a coder.
The technology behind the claim
What makes this possible is RAAPID's Neuro-Symbolic AI: an architecture that combines the pattern-recognition strengths of large language models with the rule-based rigor of symbolic reasoning.
The neural layer reads unstructured clinical notes and surfaces every code that could plausibly apply. As RAAPID CEO Chetan Parikh described it at RISE National 2026, "We let it hallucinate, we let it go nuts." The symbolic layer then applies the brakes, running each candidate code through clinical rules, CMS guidelines, credentialing checks, and coding logic before anything reaches a human reviewer.
The result is what RAAPID calls a "glass-box" system: every recommendation comes with an evidence package showing the source sentence, the clinical logic applied, and the specific CMS criteria met. Nothing is a black-box suggestion, rather everything is audit-ready by design.
In a validation study at UPMC involving nearly 47,000 charts, the platform improved coding accuracy by 26% over the prior NLP-based approach, with approximately 91% precision and coder efficiency gains between 60% and 80%.
Why this investment is different
The Healthworx investment closes what RAAPID's leadership calls a "strategic trifecta" – institutional validation from a major technology platform, a leading health system, and a dominant health plan.
M12, Microsoft's venture fund, led the original Series A, validating the underlying technology. UPMC Enterprises followed, validating the platform's clinical rigor in one of the most demanding health system environments in the country. Now, Healthworx has validated the regulatory strategy. Payers sit closest to the audit risk. Their willingness to back a platform signals something specific about where the market is going.
"M12 validated our technology; UPMC validated our clinical rigor; and now, CareFirst validates our regulatory strategy," said Chetan Parikh. "The market is no longer looking for more codes, it is looking for defensible accuracy."
Dan Zohorsky, Managing Partner at Healthworx, framed the investment in terms of the broader transformation the firm is trying to accelerate: "RAAPID is creating a new standard of accuracy that drives immense value across the entire healthcare ecosystem."
What the capital funds
The Series A extension gives RAAPID runway to accelerate enterprise partnerships with large payers and providers, and to move from point solutions toward deeper infrastructure integration – embedding its AI natively within customers' existing EHR workflows rather than operating as a standalone tool.
RAAPID currently serves 15+ enterprise customers across health plans, provider groups, and risk-bearing entities. The goal is to become the defensibility layer those organizations rely on across their entire risk adjustment operation – retrospective coding, prospective programs, and RADV audit preparation.
The company is HITRUST certified, SOC 2 compliant, and deployed on Microsoft Azure. It was recognized by Modern Healthcare as a 2025 Best in Business Healthcare IT honoree.
For more on Healthworx and its portfolio, visit healthworxventures.com.
