Every year in the U.S., hospitals lose tens of billions of dollars to a surprisingly preventable problem: documentation errors in medical records. These errors don’t just slow down billing — they lead to denied insurance claims, impact patient safety, and even open the door to lawsuits.
Our latest WorkDone podcast episode dives deep into this issue with Dmitry Karpov, CEO & Co-Founder of WorkDone, a real-time AI compliance co-pilot for hospitals. Dmitry is a two-time Y Combinator founder, former innovation leader at Ernst & Young, and a Forbes Cloud 100 rising star. His journey spans continents, industries, and multiple successful startups.
From Magic Cards to Medical AI
Dmitry’s entrepreneurial streak started early in a small scientific town outside Moscow, where he built a thriving import business selling Magic: The Gathering cards. After studying physics at Moscow State University, he moved to the U.S. for grad school, eventually launching a social media analytics startup that was quickly acquired.
At EY, Dmitry led innovation efforts, focusing on robotic process automation (RPA) — a field that taught him how to automate complex, human-driven workflows. This expertise paved the way for his first Y Combinator-backed company, Electronique, and eventually his second: WorkDone.
Why Healthcare Documentation Is a $70B Problem
Denied insurance claims are more common than most realize — 900 million are denied each year, and roughly a quarter are due to documentation issues. Missing notes, conflicting patient details, or incomplete assessments can reduce or eliminate reimbursement, costing hospitals millions annually.
And the stakes are higher than money. Inaccurate or incomplete records can harm patients and increase legal risk. For example, if a patient’s records conflict on whether an injury is on the left or right side, it can be grounds for a lawsuit — even if the medical care itself was sound.
The AI Compliance Co-Pilot
WorkDone’s platform integrates with hospital electronic health records (EHRs) to flag potential documentation issues before they cause problems. It uses AI to interpret 6,000+ regulatory rules, catch errors in real-time, and prioritize the most critical fixes. This ensures compliance teams focus on what matters most while dramatically increasing their coverage of medical records.
Instead of replacing human reviewers, WorkDone amplifies them — enabling teams to review 3–5x more cases without additional headcount.
Breaking Into Healthcare’s Toughest Market
Selling into hospitals is notoriously slow, with long sales cycles and multiple stakeholders. Dmitry’s team targets organizations with known compliance issues (often identified through Joint Commission review findings) and runs pilots on historical patient data. By showing hospitals they can automatically find the same errors their staff uncovers manually, WorkDone earns trust — and contracts.
Why Now Is the Right Time
While hospitals have tried to automate compliance before, past solutions required endless custom scripts and manual upkeep. The leap forward came with large language models (LLMs), which can interpret unstructured data, follow complex regulatory rules, and adapt without rewriting code for each new requirement.
The challenge now? Reducing false positives and ensuring the AI’s suggestions are trusted — a key focus for Dmitry’s team.
Lessons for Founders
Dmitry’s path offers several takeaways for entrepreneurs:
- Know your customer’s pain — Hospitals already spend heavily to fix documentation problems. WorkDone offers a faster, more accurate solution to an existing budgeted need.
- Choose your investors wisely — Dmitry’s cap table is made up mostly of YC alumni and healthcare-focused angels who bring strategic value, not just capital (and Team Ignite!).
- Don’t fear long sales cycles — They’re survivable if you target the right urgent problems.
- Never stop trying — His favorite piece of advice, and one he’s lived by through two YC stints and multiple pivots.
The Future of AI in Healthcare
Dmitry believes the industry’s future will be shaped by leveraging AI to multiply the impact of scarce medical professionals. With doctor shortages and rising patient loads, AI will help clinicians and nurses serve more patients effectively while also enabling patients to take more control over their own health data.
Bottom line: WorkDone isn’t just automating compliance — it’s redefining how hospitals protect revenue, ensure patient safety, and navigate complex regulations. Dmitry’s journey is proof that with the right mix of domain expertise, timing, and grit, even the most entrenched problems can be solved.
👂🎧 Watch, listen, and follow on your favorite platform: https://tr.ee/S2ayrbx_fL
Chapters:
- 00:01Meet Dmitry Karpov
- 00:48Early Entrepreneurship in Russia
- 03:40Lessons from a Teenage Side Hustle
- 05:10From Physics to Startups 07:30 – Breaking into Corporate Innovation
- 09:40Founding Electronique 12:35 – A Strategic Exit and a New Focus
- 13:43Early Customer Development for WorkDone
- 15:50Why Denied Claims Are a $70B Problem
- 23:31Patient Lawsuits and Compliance Risks
- 27:18Manual Audits vs. AI Oversight
- 30:25Fighting False Positives in AI Compliance
- 35:27Selling into Hospitals
- 39:53Why Now Is the Right Time for Real-Time Compliance
- 43:15Running WorkDone on Dmitry’s Own Records
- 44:31Fundraising Lessons from Two YC Batches
- 46:55The First Clinic Visit
- 50:18Regulatory Acronyms and Compliance Jargon
- 51:05HIPAA vs. SOC 2