Adentris Catches the Documentation Mistakes That Cost Hospitals Billions

The Macro: Medical Documentation Is a $40 Billion Problem Hiding in Plain Sight

I have a theory about healthcare: the most expensive problems are never the ones that make headlines. Nobody writes breathless coverage about clinical documentation errors. But 250 million insurance claims get denied every year in the United States, and a staggering proportion of those denials trace back to documentation mistakes. Wrong codes. Incomplete notes. Missing attestations. Procedures documented in a way that doesn’t match payer requirements. The dollar amount attached to this problem is somewhere around $40 billion annually, and most of it is completely preventable.

The current solution is manual auditing. Hospitals employ teams of clinical documentation improvement (CDI) specialists and coding compliance officers who review charts after the fact. They pull records, check for errors, send queries back to physicians, and try to fix things before claims get submitted. It’s slow, expensive, and retrospective. By the time a CDI specialist catches an error, the physician who made it has already moved on to the next patient and may not remember the clinical details needed to correct the note.

The retrospective model also means hospitals only audit a fraction of their charts. A typical CDI team reviews maybe 20 to 30 percent of inpatient records. The rest go out the door unchecked. Some of those records are fine. Some contain errors that won’t get caught until a claim is denied weeks or months later, triggering an appeal process that costs the hospital more money and more staff time.

AI-powered chart review has been on the roadmap for health IT companies for years. Epic has dabbled in it. 3M’s coding products use NLP to suggest codes. But most of these tools operate after documentation is complete. They’re faster auditors, not real-time assistants. The gap in the market is a tool that sits alongside the physician during documentation and catches errors as they happen, before the chart is closed and the claim is generated.

The Micro: Three Founders Who’ve Already Built and Scaled AI Companies

Adentris is an AI compliance copilot for hospitals. The product integrates with electronic health records and monitors clinical documentation in real time, flagging errors, suggesting corrections, and tracking protocol adherence as physicians write their notes. Think of it as spell check for medical documentation, except instead of catching typos, it catches the kind of mistakes that lead to denied claims and compliance violations.

The founding team has serious depth. Dmitry Karpov is the CEO, previously the founder of WorkDone. Alex Odin is the CTO, a three-time founder who led AI at ManyChat when the company was doing over $140 million in annual recurring revenue. He also founded Skipp, which he scaled to $7 million ARR with $2.8 million raised, and built Lingualeo, a language learning app with over 20 million users. Sergey Yudovskiy is the CPO and a two-time YC founder who previously ran ElectroNeek (YC W20), an RPA company he scaled to roughly $5 million ARR. This is not a team learning how to build products for the first time.

They’re a four-person team out of San Francisco, part of YC’s Spring 2025 batch, and they’ve already hit $54K in ARR with $500K ARR pilots in progress. For a company this early, that traction is meaningful. Healthcare sales cycles are notoriously long. Getting to paid contracts within months of founding suggests the product is solving a pain point that buyers recognize immediately.

The product connects to major EHR platforms including Epic, Athena Health, Oracle Health, Meditech, and Veradigm. That integration list is important because EHR compatibility is often the first question hospitals ask and the first reason deals die. Adentris has a library of over 5,000 compliance rules covering CMS, Joint Commission, HIPAA, and other regulatory frameworks. The AutoFix feature generates draft corrections inside the EHR for clinician approval, which means physicians don’t have to leave their workflow to address flagged issues.

The root-cause analysis capability is where I think the product gets interesting beyond basic error detection. It identifies systemic issues behind non-compliance, like staff overload, unclear procedures, or training gaps. That turns Adentris from a documentation tool into an operational intelligence platform, which is a much bigger value proposition for hospital leadership.

The Verdict

I think Adentris is entering a market with enormous demand and relatively weak competition at the real-time layer. The existing players in clinical documentation improvement, companies like Iodine Software, Dolbey, and 3M’s HIS division, mostly operate retrospectively. They review completed charts. Adentris reviews charts as they’re being written. That’s a fundamentally different product with a fundamentally different value proposition.

The risk is the healthcare sales cycle. Even with strong initial traction, scaling in hospital IT requires navigating procurement committees, security reviews, EHR integration timelines, and the general institutional inertia that makes healthcare the slowest-moving industry in tech. A four-person startup competing for attention against vendors with established hospital relationships will need to be exceptionally good at channel partnerships or direct sales.

The founding team’s track record is the strongest argument for execution. Between the three of them, they’ve built multiple companies to millions in revenue and shipped AI products at significant scale. That experience doesn’t guarantee success, but it does mean they’ve already made most of the first-timer mistakes that kill early-stage companies.

Thirty days, I’d want to see those $500K ARR pilots converting to signed contracts. Sixty days, the question is whether the AutoFix feature is actually getting used by physicians or being ignored. Ninety days, I’d want to understand the compliance improvement metrics. If hospitals using Adentris see measurably fewer claim denials, the ROI story writes itself. The market is massive, the product is differentiated, and the team has the chops. This one is worth tracking closely.