Can AI replicate an army of associates? These lawyers are betting their new firm on it

Sam Shaddox and Matt Souza have spent years on the inside of big-time legal work, as attorneys at a major Seattle firm and later as general counsels at tech companies. They’ve watched as law firms charge startup clients a fortune for work they believed AI could do faster and cheaper.

Talairis Law Group is their answer. The Seattle-based firm, launching this week, is built around the idea that AI can handle much of the work that associates at big law firms have traditionally done — and that startups shouldn’t have to pay big law prices for it.

“It’s a startup for this AI moment,” Shaddox said, “and it’s the startup that we all need and the Seattle startup scene needs. We’ve been on the other side of the aisle, and now it’s time for us to make a mark.”

The idea isn’t unique. Venture capital has poured into AI-native law firms over the past couple of years, with players like Crosby, Manifest OS, Eudia and Lawhive raising hundreds of millions of dollars combined. But Shaddox and Souza say those firms have each picked a single practice area — contract review, immigration, M&A diligence — leaving a gap nobody has filled.

“They’re all picking one lane,” Souza said, “and there’s not an AI-powered law firm that you can rely on to help you with your day-to-day as things come up, helping to pilot your ship.”

The founders: Shaddox and Souza were both partner-track attorneys at Perkins Coie, the prominent Seattle-based law firm, before moving in-house at Seattle-area tech companies.

Shaddox went on to legal roles at Big Fish Games and OfferUp before serving as general counsel at SeekOut, the AI-powered talent intelligence company. Souza was senior counsel at Zillow before becoming general counsel at Wrapbook, the entertainment payroll and financial platform.

It was that in-house experience, they say, that made the problem impossible to ignore.

“We were getting billed out the ears for work that — as we were adopting AI in-house — we saw law firms were not doing, or not doing it very well,” Souza said. “The whole economic model of law firms is broken. And so that’s where we started.”

How it works: Talairis is built around what the founders call a four-layer architecture:

• At the base is a large language model — the AI engine.
• On top of that sits what they call an agentic layer, with more than 100 purpose-built AI agents covering the range of legal tasks a startup might need.
• Above that is what they call the “client genome” — a stored profile of each client’s business, risk tolerance, contracts and operating history, so advice is never generic.
• And at the top are Shaddox and Souza themselves, reviewing and signing off on every deliverable.

“You’re not getting one-off advice that doesn’t know what your company is or does or how it thinks and operates,” Shaddox said. “You’re getting bespoke outcomes.”

In practice: As an example, Shaddox and Souza point to SAFEs: simple agreements for future equity, a common bridge financing tool for startups. First-time founders often try to handle them on their own, or bring in outside counsel at $1,500 an hour. Either way, manually working through the notes, side letters and cap table implications is painful and error-prone.

Talairis has built an agent specifically for it. Send them a SAFE, they say, and you get back more than a legal opinion.

“They don’t just get back, ‘Hey, here’s our thoughts on this convertible note’ — anybody can do that,” Shaddox said. “Instead, they get back a fully built-out cap table that incorporates the latest note, incorporates the side letter terms, and shows how that’s going to flow through their next financing.”

The pitch to startups: The firm is launching with paying customers, though Shaddox and Souza aren’t naming them yet. Talairis is bootstrapped and it’s just the two lawyers for now.

• Pricing: Shaddox says their hourly rate runs roughly half that of a typical big law attorney, and that the AI multiplies output enough that the effective cost to clients is a fraction of what they’d pay elsewhere.
• Privacy: On the question of whether client data is being used to train AI models — a real concern for startups sharing sensitive legal documents — Shaddox is direct: “The answer is no. Your data is never used to train a model.” Talairis has built confidentiality and attorney-client privilege protections into its architecture from the ground up.

The launch comes the same week Anthropic released Claude for Legal, a suite of more than 20 new connectors and 12 practice-area plugins aimed at bringing AI tools to law firms and in-house legal teams. Shaddox sees the timing as validation.

“Claude for Legal and any other LLM is a base layer,” he said. “Our unique approach is what sits on top: a law firm with elite attorneys, significant proprietary enhancements, per-client scoping, privilege protections, and the agentic architecture a generic plugin lacks. That’s what turns an out-of-the-box LLM into the best possible legal counsel for startups.”

Source link