Small Law Firms Without Engineers - AI-Enabled Practice Models

AI-native law firm is both real and a buzzword. And the version most firms are copying won't work for them. Philip Young built Garfield AI from scratch. A fully SRA-regulated law firm in the UK where the AI does the legal work, with solicitor oversight, to handle small debt claims. A platform that other firms and accountancy practices now license. It took two years, a physicist and former data scientist as a co-founder, and eight months of regulatory back-and-forth before a single claim went through. J.P. Mohler co-founded General Legal, a YC-backed fullstack AI law firm that turns commercial contracts in hours, flat fee, lawyers on Slack. Engineers make up a significant part of their team, and a serious share of their revenue goes back into the tech stack. Their clients are not buying a platform, but rather legal work that happens to be built on one. Both of them, when you look closely, made the same foundational decision: they were willing to become half a tech company. That is what made everything else possible. About a month ago, Zack Shapiro showed 7.5 million people a completely different path. Two-person firm, zero engineers, no proprietary platform. Custom Claude Skills that encode his judgment, his frameworks, and his voice. His firm managed to punch way above their weight. Zack’s post went viral because it gave people permission. You do not have to become Philip or J.P. to build something that works differently. But his model raises a question for larger (still small) firms, because it scales perfectly to one person, a single practitioner who is the foundation. They know where every document lives, control every workflow, and are the single brain the tools are built around. The moment you go to a team of a larger headcount, the questions change entirely. Whose judgment do you encode? Who maintains it when it goes stale? How do you standardize across a group of people who all work slightly differently? You cannot skip to Claude Skills without the operational layer underneath them. But enough with my analysis. I want to hear from firms that are building somewhere between Zack on one end and JP and Philip on the other. Without fulltime engineers or developing their own platform. The ones operating on a combination of existing tools stitched together intelligently, working dramatically faster and better internally than anyone looking at them from the outside would expect. I know these firms exist. And most of them are not talking about it publicly. If that is you -what made you choose this path? Did you start from scratch or transform an existing practice? And what’s the hardest problem you haven’t solved yet? --- 👉 Send me a message. I suspect a lot of interesting versions of this are happening quietly. Lev Loukhton, Kyle Bahr, Damien Charlotin, Jamie Tso, Jiyun Hyo, Helen Fan, Kaj Rozga, Anastasia Boyko, Ben Chiriboga, Kevin Keller, Nate Kostelnik, Tom Rice, Alex Herrity