On a Tuesday in late April, Neal Cheskis got something he hadn't expected: a resolution to a rejected insurance claim that had been languishing in his practice's billing queue. The culprit wasn't a new hire or a billing specialist finally getting around to it. It was an AI system run by a San Francisco startup so small that when you email them, you're probably talking directly to one of the founders.
Taiga—officially Bayes AI, Inc., though no one seems to call it that—went live on April 29 with a pitch that sounds almost reckless for a team of two: 98% first-pass claim acceptance, plus automatic appeals on every single denial. No exceptions, no extra fees, no physician having to decide whether chasing a $200 rejection is worth the administrative calories.
It's an audacious entry into what has become, rather suddenly, a very noisy room.
The AI Revenue Cycle Gold Rush
The past few months have seen an unusual concentration of moves in the medical billing space. Waystar announced "agentic AI" for autonomous revenue cycles in February. Ventra Health rolled out vCision that same month. RethinkBH launched BillAI for behavioral health in April. The timing isn't coincidental—investors have caught the scent of automation in healthcare's most notoriously manual corner.
William Blair published equity research in May with a title that doubled as thesis statement: "The Growing Importance of AI in the Revenue Cycle Management." Commure, which owns the billing platform Athelas, raised $70 million in May at a $7 billion valuation. Adonis closed a $40 million Series C in March. Tebra pulled in $250 million in combined equity and debt last December, partly to fund AI expansion across documentation and billing.
Everyone, it seems, wants a piece of the revenue cycle. The question for Nanda Guntupalli and Adam Wax, Taiga's co-founders, is whether a two-person operation can carve out something defensible—or whether they'll get swallowed by better-capitalized competitors before they find scale.
What Makes This Different (Maybe)
Taiga's product handles the standard billing gauntlet: eligibility checks, prior authorizations, AI-generated diagnosis and procedure codes, claim scrubbing, submission, appeals, patient statements. It plugs into major electronic health record systems—athenahealth, eClinicalWorks, NextGen, Epic, AdvancedMD—without needing IT support from the practice, which matters when your customer is a solo physician who can barely keep up with charting.
But here's where the founders think they've found an edge: physicians stay in the loop. The AI drafts ICD-10 and CPT codes, but doctors review and sign off before anything goes out the door. It's a hybrid play, trying to split the difference between full automation and the clinical accountability that keeps practices from getting audited into oblivion.
The go-live timeline is supposedly two weeks. Coverage spans commercial payers, Medicare, Medicaid, TRICARE. Cheskis, who runs a family therapy practice, told the company that Taiga resolved his stuck claim in about a week—though one data point doesn't make a trend.
Denials on Autopilot
Perhaps the most interesting gambit is the automatic appeals policy. Every denial triggers a workflow, no questions asked. Taiga claims its system includes a "voice agent" that can chase down missing authorizations and reprocess claims through electronic remittance channels.
The company illustrates this with a scenario: a claim denied for code CO-197 (missing authorization) gets auto-appealed and resubmitted. Whether that scales operationally with two founders and whatever handful of contractors they might have is, to put it mildly, an open question. Managed billing services aren't exactly known for their operational leverage. You can automate a lot, but someone still has to manage edge cases, talk to insurers, handle practice-specific quirks.
Still, the positioning is sharp. Denials, Taiga argues, should be baseline service—not an upsell or an optional add-on.
The Economics of Small Practice Billing
Taiga doesn't publish pricing, which is often a tell. It's a managed service, not self-serve software, with custom pricing based on claim volume. That contrasts with the typical models in the space: athenahealth bundles EHR and billing at 3–7% of net collections plus implementation fees. Tebra charges $49–$799 per provider per month for software subscriptions.
Taiga positions itself as EHR-agnostic and free of percentage-based fees. For practices already invested in a particular EHR—or allergic to giving up a slice of every dollar collected—that could be appealing. But without transparent pricing, it's hard for a prospect to do the math upfront.
Founders Who Lived the Problem
Both Guntupalli and Wax grew up watching their parents run small medical practices. According to the company's launch materials, that proximity to the daily grind of billing chaos informed their conviction that automation could finally break the cycle.
They went through Y Combinator's Spring 2026 batch and raised $125,000 from the accelerator in March, per Dealroom data. YC lists the founding team at two; LinkedIn pegs the company at 2–10 employees as of late May, which likely includes contractors or early hires who haven't yet formalized their roles. Either way, it's lean.
The Wedge Question
Taiga's bet seems clear enough: go after independent practices that feel underserved by enterprise RCM vendors and don't want to overhaul their existing tech stack. Make denials someone else's default problem. Let physicians keep clinical oversight without drowning in paperwork.
What's less clear is whether that wedge is wide enough—and whether a managed service model can scale without the kind of operational infrastructure that usually requires more than two people and a YC check. Software scales nicely. Managed services... not so much, at least not without finding leverage in ops or hiring aggressively.
For now, the company is keeping it direct: a "Talk to a founder" booking link, an email address ([email protected]) that presumably goes straight to Guntupalli or Wax. It's intimate, maybe out of necessity as much as strategy.
The AI revenue cycle wave is real, and the capital is following. Whether a two-founder startup can ride it—or get pulled under—depends on whether Taiga can deliver on that 98% claim rate and auto-appeal promise at scale. The early customers seem pleased. The next hundred will be the real test.
