The average dental clinic burns more than 160 hours a month on insurance paperwork. Toothy AI, a Winter 2025 Y Combinator company, is handing that work to AI agents that call payers, file claims, and chase the money down.
There are two businesses inside every dental office. The first one fixes teeth. The second one argues with insurance companies about who pays for it. The first is why anyone goes to dental school. The second is the reason the front desk looks tired by 10 a.m., stuck on hold, reading policy numbers into a phone. Toothy AI, a company from Y Combinator's Winter 2025 batch, is a bet that the entire second business can be handed to software.
The founders put a number on the problem: the average dental clinic spends more than 160 hours a month on insurance tasks - verifying patient policies, filing claims, correcting denials, posting payments. That is roughly one full-time employee whose whole job is paperwork and hold music. Across the industry, Toothy estimates the waste at $13.2 billion a year, with about 80% of practices reporting financial strain tied to how badly the revenue cycle runs.
Toothy describes itself less as an AI company and more as an AI insurance coordinator for dental clinics. The distinction matters, because the buyer is a practice owner who does not want to purchase artificial intelligence - they want to stop losing money to unverified coverage and unpaid claims. Under the hood, Toothy runs HIPAA-compliant AI agents that handle the full revenue cycle: verifying primary and secondary coverage by calling payers and navigating their online portals, then writing the benefit details straight into the practice management system up to two weeks before the patient walks in.
The headline capability is the one that sounds most like science fiction and turns out to be the most tedious in practice: a voice AI that places the call to the insurance company and survives the hold queue. From there the agents build and submit claims with 50-plus clinical codes, correct denials, draft appeal letters, and post the payments when the insurance checks come in.
Voice AI and portal bots confirm coverage, written to the PMS 2 weeks early.
Clean claims submitted with 50+ clinical codes.
Denials corrected, appeal letters drafted automatically.
Insurance payments reconciled and posted back to the practice.
Figure 1 - Toothy's revenue cycle, verification through payment.
The customer is the dental practice - single-location offices and growing groups that would rather not staff a full billing department. Toothy's pitch to them is unusually concrete for an AI product. It publishes outcomes: one clinic reported its collection ratio climbing from 50.8% to 97.2%, meaning it had been collecting barely half of what it was owed. Another cleared $119,172 in aging accounts receivable and closed 94 old claims. The company says early customers cut their back-office teams by 50-75% within weeks, and estimates practices save anywhere from 80 to 240-plus hours a month depending on size.
Figure 2 - Reported customer result. Individual outcomes vary by practice and payer mix.
Most of the revenue a practice "loses" isn't fraud or bad debt. It's claims that were never verified, never followed up, or written off because nobody had time to appeal. Automating the follow-up is where the collection ratio climbs.
Toothy AI was founded in 2024 by Johnny Chen and Matthew Kerrigan, both second-time founders. Chen, the CEO, previously built one of the earliest AI B2B cold-email tools - reaching $1 million in annual recurring revenue using GPT-J, before ChatGPT existed - and earlier worked as a software engineer at Ford Motor's R&D lab. Kerrigan came from the technical side, having built Salezilla, an LLM-based cold-email product, and held a machine-learning role at a startup later acquired by Nordstrom.
The through-line is a pair of engineers who already know how to sell automation to businesses, now pointed at a market nobody brags about at dinner parties. Dental insurance billing is not a topic that trends. That is arguably the point - the boring, high-friction corners of an industry are where a working AI agent has the clearest job to do.
Toothy sells the service as a subscription. There is a per-provider plan with unlimited verifications priced monthly per dentist, and usage-based bundles that come with a monthly verification allowance plus overage options. Larger practices get custom quotes based on size, appointment volume, and payer mix. Alongside the automation, customers get real-time dashboards and daily reports tracking verifications, billing, collections, and aging - so a practice owner can check the back office once a day instead of living inside it.
In the market, Toothy sits against two incumbents: outsourced dental billing firms and eligibility tools that still lean on human labor, and the in-house biller a practice would otherwise hire. Its differentiation is the degree of autonomy - voice AI making the actual payer calls and agents running claims end to end, rather than software that hands a worklist back to a person. The company operates out of San Francisco and worked with YC partner Jared Friedman during the W25 batch.
Toothy doesn't lead with "AI." It leads with the outcome a dentist actually wants - getting paid faster with fewer people on the phone. The technology is the how, not the pitch.
Whether Toothy AI grows into a category or stays a promising niche will depend on things that are hard to see from the outside: how reliably the voice agents clear payer phone trees, how well the claims hold up across thousands of plans, and whether the compliance story satisfies practices handling protected health data. What is clear is the shape of the bet. Take the single most dreaded, most repetitive job in a service business, and let an AI agent do all of it. If it works in dentistry, the same template is waiting in every other clinic that fights insurance for a living.
Toothy AI had not published an official YouTube channel or product-demo video at the time of writing. A live product walkthrough can be booked directly through the demo link on toothy.ai.