A handshake in a freshman dorm, four years late
Walk into Sandstone's Brooklyn office at 11pm and you might find the founders just getting back from dinner - and heading straight to their desks. Sequoia noticed. Most investors talk about energy. This was the kind you could trip over.
Nick Fleisher is the co-founder and CEO of Sandstone, the company he calls the first platform for AI-native legal departments. The pitch is deceptively plain: engineers got Jira and Linear and a decade of tools designed around how they actually work. The lawyers sitting one floor away got Slack pings, procurement tickets, and a SharePoint folder nobody can find. Sandstone is the attempt to close that gap - to take the playbooks, instincts, and judgment trapped in people's heads and turn them into something a company can actually use.
"You think about how engineers have JIRA and Linear. On the legal side, that doesn't really exist."
The origin is older than the company. Fleisher met Liam Germain - one of his eventual co-founders - in a Penn Class of 2022 Facebook group, of all places, where the two bonded over shared interests including fishing. They became first-year roommates at Hill College House. Somewhere in those early months they made what Fleisher calls a "handshake deal" to build a startup together one day. Then they tried to do it immediately. They started building mobile apps. They talked about dropping out. The apps failed. Both of them, sensibly, finished their degrees.
That failure matters, because it explains the temperament. By the time Fleisher graduated from Wharton in 2022, he had shipped more than ten mobile apps across AI and consumer. None of them made him famous. All of them taught him how to start over, which turns out to be the only skill that compounds.
How to lead a practice nobody assigned you
Fleisher joined McKinsey & Company in New York as a consultant advising some of the largest companies in the world. By his own account he reached engagement manager in roughly eighteen months by "hacking the system" - specializing early in legal tech instead of waiting his turn in the traditional promotion ladder. He then, in his words, "accidentally found himself leading the firm's legal tech and AI strategy practice."
It is a strange thing to fall into a specialty, but the timing was perfect. He spent his entire McKinsey career serving general counsels and mid-market in-house legal teams, watching from the inside as AI went from a slide-deck buzzword to something that could actually do the work. He also picked up a management habit there that he'd later install at his own company: he watched senior partners openly ask junior people for feedback, and decided that was how a real team should run.
"The biggest risk in legal AI isn't necessarily the tools themselves. It's how much automation people allow."
That line is the whole philosophy in miniature. Fleisher is not selling a robot lawyer. He is selling leverage for understaffed teams - a system that learns a company's legal intelligence and helps execute the repetitive parts, while leaving the judgment where it belongs. Law is, as Sequoia put it, an almost perfect fit for large language models: text in, text out, expensive human labor doing repeatable work. The trick is knowing where to stop.
An MVP in weeks, a believer from 2019
In 2025, Fleisher left to start Sandstone with two people who completed the triangle. Jarryd Strydom is the rare combination of attorney and builder - an in-house lawyer who taught himself to code while still practicing, and who had lived the exact pain Sandstone targets, buried under "Slack pings, procurement tickets, and redlines." Liam Germain, the freshman roommate and an economics graduate, finally made good on the handshake.
They moved fast. The team built a working MVP in weeks and put it in front of customers immediately. That speed, more than any deck, is what drew Sequoia. The recruitment had been going on for months - Sequoia's talent team had introduced partner Bogomil Balkansky to Fleisher, and even a competing cybersecurity CEO had tried to recruit him - but the detail that "may have sealed the deal," Fleisher says, was almost cosmic.
Back in 2019, Balkansky had sketched a nearly identical legal-tech concept after watching a ten-person team at VMware manually pull terms out of contracts. He'd written a product requirements document and shelved it. When he met Fleisher, he dug it out and sent it over. Two people, six years apart, had drawn the same picture. The seed round - $10M, led by Sequoia, with more than twenty general counsels, law firm partners, and customers joining in - closed and was announced in January 2026.
"This is not our first attempt. It's never too early to build, and it's never too late."
No DMs, no ego, ship something
Run Sandstone the way Fleisher does and you'll notice the private messages are missing. Most communication happens in public Slack channels by design - "no ego," shared context, everyone seeing the same thing. It's the McKinsey-partner habit, scaled down and turned up: the person with the most context shouldn't be the person with the most secrets. He likes to hire people who can ship and who code to some extent, regardless of the title on their resume.
Two months after launch, the company - around two to three dozen people in Brooklyn - was running more than a hundred customer demos a week. Fleisher has called the current moment a "seller's paradise" for in-house legal AI, the rare window when the market wants the thing before you've finished building it. The goal he keeps returning to is small enough to fit on a sticky note and large enough to organize a company around.
"Our goal is that if you're a lawyer at a company, the first thing you open in the morning is Sandstone."
It's a modest sentence with a heavy ask behind it. Being the first tab someone opens is the hardest real estate in software - you have to be useful before coffee, every single day. But it's a fitting target for someone whose entire arc is about persistence: the failed dorm-room apps, the ten that followed, the specialty he fell into and then ran, the handshake he honored four years late. Fleisher isn't betting that AI will replace lawyers. He's betting that the lawyers who keep their judgment and hand the busywork to a machine will quietly run circles around the ones who don't. And he'd like to sell them the machine.