AI-Driven Healthcare Claims Advocacy
January 2026
The U.S. healthcare system denies roughly 11.8% of all initial claims — up from 10.2% just a few years ago. That translates to $262 billion in denied claims annually, based on 2016 data when total claims volume was $3 trillion. With healthcare spending now exceeding $4.5 trillion, the denial problem has only grown.
McKinsey's research on healthcare service operations points to a fundamental issue: payers and providers operate in adversarial systems with misaligned incentives. The complexity is intentional. Prior authorization requirements have exploded — 73% of providers reported increases in 2024 alone. The result? Patients caught in the middle, often abandoning legitimate claims rather than navigating the appeals labyrinth.
The human cost is harder to quantify but no less real. A denied claim for a necessary procedure doesn't just mean lost revenue for a hospital — it means a patient delaying care, paying out of pocket, or foregoing treatment entirely. The Kaiser Family Foundation found that surprise medical bills affect roughly 1 in 5 insured adults.
Where AI Changes the Equation
Traditional claims advocacy relies on human experts billing hourly — effective but inaccessible to most patients. AI changes the economics:
- Document analysis: LLMs can parse Explanation of Benefits (EOB) documents, policy language, and denial letters in seconds
- Appeal generation: Drafting medically necessary letters and state-specific appeals based on clinical guidelines
- Process automation: Tracking deadlines, submitting paperwork, and following up with payers
McKinsey estimates that generative AI could automate 30% of hours worked in healthcare by 2030. The highest-value applications aren't replacing doctors — they're eliminating administrative friction that harms patients.
What We're Building
At Do Little Lab, we're building an AI-native platform to help patients contest wrongful denials. The goal isn't to replace human advocates — it's to make their expertise scalable and accessible.
Early results are promising: patients who would have paid thousands out-of-pocket are getting overturns. The system still requires human judgment for complex cases, but AI handles the legwork that previously made advocacy uneconomical for smaller claims.
The larger bet is that making denials easier to fight will change payer behavior. If wrongful denials are consistently appealed and overturned, the business case for denying care in bad faith erodes.
—
Building Do Little Lab. Previously: SRE, engineering manager, humanitarian relief organizer.