Third Chair Reads Every Ad on the Internet - and Sends the Bill

A record label learns that a song it owns is playing behind a car dealership's ad in Ohio, a fitness brand's reel in Miami, and a supplement spot nobody can quite place. In the old world, someone would decide whether it was worth a lawyer's hourly rate to chase any of it. Usually it wasn't. The infringement stayed, and the money stayed on the table. Third Chair was built to change that arithmetic.

The San Francisco company, part of Y Combinator's Spring 2025 (X25) batch, describes itself in plain terms: an AI-native IP enforcement agent. It finds unauthorized uses of a rightsholder's copyrights, trademarks, and name-image-likeness across the web, tracks down whoever is responsible, and drafts the paperwork to get them to pay or license properly. The AI does the searching and the first-draft writing. Human lawyers stay in the loop to review and sign.

That quote, which the company puts on its own site, is the whole opportunity in one sentence. The incumbent product in this category is a spreadsheet. Someone overseas labels the infringements, ships a CSV, and the rightsholder is left to do everything that actually matters - figure out who to contact, write the letter, follow up, close the deal. Third Chair's bet is that the finished job, not the raw data, is the product worth paying for.

There's a reason that gap has stayed open for so long. Detection alone was never the hard part; plenty of vendors could tell a label its catalog was showing up in places it shouldn't. The hard part was the fifty steps after detection, each of which needed a human who understood both the legal posture and the specific ad in front of them. That work didn't scale, so it didn't happen. What's changed is that AI agents can now carry a matter from a flagged clip to a drafted demand without a person touching every rung of the ladder - and the person who does step in arrives at the end, with the context already assembled.

01 / What it doesThree agents, one workflow

The product breaks into three moving parts. A browser agent monitors platforms for uses of a client's IP. An outreach agent finds the business behind an unauthorized ad, verifies that an email will actually land, and handles follow-ups. A contracts agent drafts demand letters, pulling in the context the other two gathered, for a lawyer to review. Strung together, they turn a scattered detection problem into something closer to a pipeline.

Fig. 1 - The monitor-enforce-license loop, as described on Third Chair's site.

02 / Who uses itStarting where the rights are messiest

Third Chair started with music, and for a reason: music has some of the most tangled rights and the most motivated owners. The company says it now works with more than 60% of the recorded music industry, including Universal Music Group. Music distributor Symphonic Distribution and the startup Duetti are named users. Smaller shops like Fat Coda Studio and New 11 Records appear on its site too.

The revenue figure is worth reading carefully. Third Chair says its customers generated more than $1M in the last six months of 2025 using the platform. That's money the rightsholders collected, not the startup's own top line - but it's the number that matters to the pitch, because it reframes what a legal tool is for.

03 / The problemEnforcement as a profit center

Most legal departments are cost centers. You spend money to avoid worse outcomes. IP enforcement has traditionally sat squarely in that bucket - a defensive expense, triaged down to only the biggest offenders because chasing the rest didn't pencil out. Third Chair's argument is that AI drops the cost of the search close enough to zero that the payout on the other end becomes margin. Change what it costs to look, and the whole economics flip from defense to income.

04 / The modelGetting paid to get others paid

How Third Chair itself makes money follows from the pitch. It's a B2B software company - rightsholders pay for access to the agents and the lawyer network - but the more interesting piece is that its incentives line up with recovery. When enforcement is framed as a revenue stream rather than a compliance chore, the vendor that helps you collect can share in the upside instead of billing flat hours regardless of outcome. That alignment is unusual in legal services, where the meter runs whether or not you win, and it's part of why a distributor's head of sync would describe the relationship as a revenue stream "we're excited to build" rather than a line item to minimize.

It also explains the choice to keep lawyers on staff. A pure software play could ship detections and drafts and stop there. But a demand letter is only as good as the credibility behind it, and credibility in this world still runs through licensed attorneys. By putting counsel inside the loop rather than leaving clients to find their own, Third Chair sells something closer to an outcome than a tool - which is harder to build and harder to copy.

05 / The differenceWhy not just contract-review AI?

The crowded lane in legal AI is contract review and legal research - reading documents faster. Third Chair went sideways into the workflows that generate money rather than save time. It also refuses the pure-software shortcut: rather than sell a dashboard and wish clients luck, it keeps lawyers inside the loop, which is what lets a demand letter carry real weight. The name fits the strategy. In a courtroom, the "third chair" is the junior attorney doing the unglamorous prep behind the lead lawyers. The company named itself after the work nobody wants and then automated it.

06 / The peopleFounders who keep selling to music

CEO Yoav Zimmerman previously founded Trendpop (YC W21), an analytics company he scaled past $1M ARR selling to the likes of Universal Music Group and MrBeast before exiting. Earlier he worked on machine learning at Google. His co-founder and CTO, Shourya Lala, previously built Fello (YC W22) to more than two million users and spent time at BlackRock. Between them, that's three prior YC batches and a founder who has already sold into the exact industry Third Chair is chasing - useful when your first customers are labels who don't hand trust out freely.

The mission line is a bet on a specific future - solo founders with enormous reach and no in-house counsel. Whether that world arrives on schedule is an open question. The near-term version, though, is already running: labels with plenty of lawyers and not nearly enough hours, using software to cover ground people never could.

07 / The marketWhere it goes next

Third Chair began with music copyright, but the same three agents point at a wider target. The company has signaled plans to expand into name, image, and likeness - the NIL rights that matter to athletes and creators - as well as trademarks. Anyone who owns something that gets used without a check is a potential customer. The company is backed by Y Combinator and FundersClub, with $2.3M raised as of its 2025 seed. It's a small team, hiring both engineers and, tellingly, IP enforcement counsel to sit alongside the code.

The uncomfortable truth underneath all of it: most infringement goes unpunished not because owners don't care, but because caring costs more than it returns. Third Chair didn't set out to argue that infringement is wrong. It set out to make chasing it cheap enough to be worth doing - and to let the people who own the rights keep the difference.