Kristina Subbotina left Cooley and built a startup-focused law firm on TikTok—then turned her practice into an AI-powered legal operating system called Lexsy. The pivot helped her grow Lawlace to more than $1.3 million in revenue in two years, and later launch
When Kristina Subbotina left Cooley—a Big Law firm where she’d spent nearly a decade as a corporate startup lawyer—she didn’t picture herself becoming the kind of attorney founders would want to follow.
She thought she was moving toward investing. “I decided I wanted to join a VC fund as an investor,” she said in an as-told-to conversation that was edited for length and clarity. But the plan bent fast.
To attract potential portfolio companies, she began posting legal tips for startups on social media. It worked—until the messages stopped asking for her investment and started asking for her help. “All of the founders who reached out to me wanted legal support, not an investor,” she said. Some even offered equity in addition to cash for her services.
Mentors told her to stop chasing what she thought would succeed and to respond to what the world needed from her. Subbotina took that advice and started her own startup law firm called Lawlace. In two years, it grew to more than $1.3 million in revenue. She’s now expanding the business by productizing her expertise—moving from services to software through Lexsy. an AI-powered legal operating system.
In Big Law, she said, there was a stigma around social media—the notion that a top lawyer dancing on TikTok would be cringe. Subbotina chose the opposite approach anyway, building her outreach on two principles: “meet people where they are and provide value.”
Her videos leaned into short, fast storytelling. She told “real legal horror stories” in clips that were sometimes danced-through and sometimes comedic, each one designed to make founders pause long enough to spot a problem they might be facing.
A 20-second story, she said, could trigger the kind of realization founders don’t forget: “Wait, we never signed IP assignments,” or “We don’t have vesting,” or “Did we file an 83(b) on time?” Those clips went viral.
The numbers became impossible to ignore. Her videos received “over 5 million monthly views,” and Mark Cuban reposted her content. Instead of legal knowledge living only inside partner meetings, it reached founders where their attention already was. Subbotina said the outcome was straightforward: people watched her videos. avoided costly legal mistakes before they could afford her. and eventually turned into clients.
She frames content as a trust-builder at scale. “Law is a relationship of trust, and content is a way to build trust at scale,” she said. She describes her social media output as an “evergreen investment” that continues bringing new clients even when she isn’t actively chasing acquisition.
The next shift came from the friction of scaling a service.
As with many founders who start in client work, she said scaling Lawlace was hard because legal work often lives “only in lawyers’ heads”—manual, slow, and difficult to expand.
She didn’t want a traditional law firm. So she began automating workflows early. By late 2025, she said, it was clear AI agents could handle work that amounted to “the job of a junior associate.” That realization pushed her toward a new iteration of the business: Lexsy.
Lexsy positions itself as an AI-powered legal operating system for companies. In her model, clients buy a subscription and run their entire legal life on the platform. She said everything is protected by attorney-client privilege. and the platform runs “on autopilot with a human lawyer in the loop. ” aiming to reduce friction for founders.
Her company is built to be AI-native. She described an “army of AI agents” that take a first stab at tasks, with senior team members handling strategy and review. That structure, she said, is how she plans to grow quickly without adding headcount.
The human team is small: five people total—Subbotina herself, a head of product and operations, and three software engineers. She also said she relies on a handful of part-time support for a podcast, AI architecture, and social media.
To fund the transition, Subbotina raised a relatively modest amount with very specific goals.
Last year, she closed a fundraising round of $650,000. “It’s nothing by Silicon Valley standards,” she said, but she framed her approach as deliberate because Lawlace already had revenue and was profitable. She was “intentionally picky” about what she accepted and why.
Her first priority was time to build. She wanted enough money to “stop sales for six months” so she could develop Lexsy. She had already bootstrapped and tested the platform before raising, and she said the funding was needed to build the platform and agents and to hire the team.
She also set a hard rule for the right investors. If she felt an investor couldn’t help beyond capital—such as “media support or customer introductions”—she wouldn’t take the money.
That funding gave her room to take a risk that could have failed. Instead, she said Lexsy emerged from stealth in June 2026. It went from quietly building behind the scenes to launching in public, with $372,000 in annual recurring revenue.
The recurring revenue came from clients who migrated from the law firm, she said, along with word-of-mouth growth and her existing social following.
The story she tells service-business owners is blunt: productize yourself.
“Keep the human touch for what matters, but automate execution,” Subbotina said. She adds that you don’t have to be technical and you don’t have to raise venture money to apply the principle—if you provide value, the fundamentals still hold.
For Subbotina, the through-line is trust—first earned through short-form videos that helped founders avoid expensive mistakes, then translated into a platform designed to deliver legal help with AI agents in the driver’s seat and a human lawyer stepping in when it matters.
Lexsy Lawlace Kristina Subbotina startup law AI legal TikTok Cooley Mark Cuban fundraising annual recurring revenue attorney-client privilege legal automation