The medical assistant job wasn't what Meng Fei Shen expected after leaving Uber. Spending months in a California rheumatology clinic, he watched staff drown in a peculiar kind of administrative quicksand: prior authorizations for biologic drugs, those complex medications that can run $80,000 to $150,000+ per patient annually and require navigating what seemed like an endless maze of insurance portals.
That immersion—part career pivot, part field research—became the foundation for Ruma Care, a startup that emerged from Y Combinator's Winter 2026 batch with a narrow but potentially lucrative thesis. While most prior authorization platforms cast a wide net across all of healthcare, Shen and co-founder Christina Huang are betting that infusion clinics managing specialty medications represent a neglected automation opportunity. The paperwork burdens are worse, the financial stakes higher, and the workflows distinct enough to justify purpose-built software.
Perhaps more than the founders initially anticipated, early traction suggests they might be onto something. Ruma claims its platform has reduced denial rates by 50% to 55%—figures that vary slightly between the company's Y Combinator profile and LinkedIn announcements—and cut staff time per authorization by a factor of 20 across clinics in California and Nevada where it's currently operating. The usual caveats apply: these are self-reported figures from a company barely six months old, with no third-party validation published yet.
Still, the problem Ruma is attacking is real enough that even large incumbents are scrambling to automate it.
Seventy Portals and Counting
The core issue isn't just volume. Biologics, used to treat conditions like rheumatoid arthritis and Crohn's disease, require clinics to navigate what Ruma describes as more than 70 separate online portals across different payers. Each comes with its own approval criteria, forms, and submission quirks. For buy-and-bill infusion centers—where clinics purchase expensive drugs upfront and seek reimbursement—a denied authorization doesn't merely delay care. It can mean absorbing the cost of medication already sitting in inventory.
Ruma's software pulls data directly from electronic health records to auto-populate these payer-specific forms, extracting the precise criteria each insurer demands. It handles copay assistance enrollment using the same integrated data and flags potential reimbursement issues before denials materialize. The company says its models learn from denial patterns over time, theoretically improving submission quality as they process more cases—though that's a claim several competitors also make, and independent validation of these learning systems remains scarce across the industry.
On January 21, the startup launched a free Formulary Navigator tool. Providers can search coverage rules and prior auth requirements by plan, medication, and diagnosis through a natural-language interface that surfaces information from official payer documents. It's a utility play, designed to reduce the upfront research burden before even starting an authorization.
An Unusual Path to Healthcare Tech
Huang's route into this problem involved her own experience navigating biologic access as a patient, combined with previous roles at Apple and building integrations at Medallion and UnitedHealthcare. Shen's decision to work as a medical assistant after stints at Uber and Walmart Data Ventures was more deliberate—an attempt to understand prior auth workflows from the inside before trying to automate them.
That ground-level perspective is evident in how they've positioned the product. Ruma is explicitly "built for rheumatology" and "designed for infusion clinics," a narrower focus than the horizontal platforms dominating the market. The San Francisco-based company lists just two employees on its Y Combinator profile. They appear to be starting deliberately small, perhaps waiting to prove the model in one specialty before expanding.
There's a certain logic to that approach. Specialty drug workflows differ meaningfully from standard prior authorizations—higher costs, different purchasing models, and payer rules that change more frequently. Whether those differences justify an entirely separate software stack is the bet investors will have to evaluate.
A Market Already Crowded With Automation
The prior authorization space is experiencing something of an AI gold rush, and Ruma is hardly alone in promising faster, smarter approvals. CoverMyMeds, now owned by McKesson, announced what it called "industry-first integrated medication access" solutions for specialty therapies on March 9, 2026. SamaCare launched its AI-driven SamaCare Plus offering last November, promising touchless prior authorizations with human oversight. Cohere Health raised $90 million in May 2025 to scale its AI platform on the payer side of the equation.
Even pharmacy benefit managers are automating aggressively. Optum Rx reported in October 2024 that its PreCheck Prior Authorization system could process certain drug PAs in a median 29 seconds, down from the 15 to 20 minutes that used to be standard. That kind of speed matters when delays translate directly into lost revenue or patients abandoning treatment.
The regulatory environment is shifting, though not always in ways that help drug-specific prior authorization. CMS's January 2024 Interoperability and Prior Authorization Final Rule mandates FHIR APIs for prior auths by 2027—but explicitly excludes drugs from those requirements. Pharmacy prior auth remains fragmented across dozens of disconnected portals, which is precisely the fragmentation Ruma is attempting to solve with one unified workflow.
The Validation Problem
Beyond a testimonial from San Mateo Rheumatology praising the platform's speed, public evidence of Ruma's deployment scale remains thin. The company hasn't disclosed customer counts, revenue figures, or pricing. No independent studies have verified the 50%-to-55% denial reduction or the 20-times staff efficiency improvement—both numbers that would be transformative if they held up at scale.
For digital health investors evaluating the current wave of prior auth automation, Ruma represents a specific wager: that vertical focus beats horizontal scale in this market, and that specialty drug workflows differ enough from standard PA to warrant dedicated tooling. It's a reasonable thesis. Whether it proves out depends on how quickly the team can demonstrate outcomes beyond early California adopters, and whether those metrics survive scrutiny from procurement committees at larger health systems.
Two employees tackling 70-plus insurance portals. It's either audacious or absurd, depending on how the next year unfolds.
