I Interviewed the Claude-Native Lawyer Whose Viral Post Broke the Internet, and What He Said Changed Everything...
Zack Shapiro’s “Claude-Native Law Firm” hit 7.5 million views. But the real breakthrough isn’t the workflow. It’s what it reveals about the future of our professional identity
Last week, Zack Shapiro published a post on X called “The Claude-Native Law Firm.” If you’re reading this newsletter, you probably saw it. Over 7.5 million people did. More than 320 comments. Thousands of reactions, and lawyers reaching out to him directly. For context, there are about 1.3 million lawyers in the United States. This wasn’t a legal tech moment. It was a cultural moment — a signal that the anxiety around AI and professional work had been waiting for someone to say: here’s a path forward, and it’s not as scary as you think.
I got Zack on the LawDroid Manifesto Podcast this week (that episode drops Monday) and the conversation went deeper than the article. What I want to share isn’t a summary of his workflow. It’s what I think his breakthrough actually means, and why it matters far beyond one lawyer’s practice.
If this sounds interesting to you, please read on…
This substack, LawDroid Manifesto, is here to keep you in the loop about the intersection of AI and the law. Please share this article with your friends and colleagues and remember to tell me what you think in the comments below.
The Real Unlock: Mindset, Not Technology
Most of the commentary around Shapiro’s post has focused on the how. The skills. The prompts. The tracked changes. All of that is impressive. But in our conversation, Zack said something that cut through the tactics: “The limiting factor is not your skill. It’s your imagination and your willingness to go for it.”
That’s not a statement about technology. That’s a statement about mindset.
For three years, the legal profession has been circling AI like it’s a dangerous animal. Cautious. Skeptical. Waiting for someone else to go first. What Shapiro did (and why it resonated with millions) is he walked into the room and said, I went first, and I’m thriving. Not surviving. Thriving. Having more fun practicing law than ever before.
The unlock isn’t Claude. It’s permission. The permission we grant ourselves to reimagine what it means to be a lawyer when the tedious, mechanical parts of the work (the formatting, the first-pass drafting, the citation checking) are handled by a machine, and what’s left is the part you actually went to law school for. The judgment. The credibility. The human relationship with the client who needs you to understand what’s really at stake, for them.
Zack put it beautifully in our interview: “It’s not how we do things. It’s how I do things.” That single shift, from institutional conformity to individual expression through AI, is the psychological breakthrough most people missed in the viral noise.
“The limiting factor is not your skill. It’s your imagination and your willingness to go for it.”
The Mindset Beneath the Method
Here’s what I’ve learned teaching law students at Suffolk University Law School and building AI legal tools for over a decade: the technology is never the hard part. The hard part is the shift in thinking that lets you use it.
Shapiro’s prompts for complex assignments average 2,000 words. He’s not chatting with AI. He’s briefing it the way a senior partner briefs a trusted associate, with context, nuance, judgment, and specificity. That skill, the ability to be precise and detailed, logical, in written instructions, is literally what “thinking like a lawyer” develops. Lawyers don’t need to learn prompt engineering. They need to recognize that they’ve been doing it their entire careers.
This is what I call the hidden superpower of the legal profession. The same discipline that makes a great contract drafter, anticipating edge cases, being specific about terms, structuring logic clearly, is exactly what can make someone effective with AI. The profession that’s most anxious about being replaced is, paradoxically, the one best equipped to lead the transformation.
In our conversation, Zack talked about how he hates learning tools. Hates dashboards and toggles. Has ADD. And yet he’s built one of the most sophisticated AI-integrated legal practices in the country. Why? Because this technology doesn’t ask you to conform to it. It conforms to you. You speak to it in plain English. You tell it how you think. And it remembers and understands.
That’s a fundamentally different relationship between a professional and their tools than anything we’ve seen before. And it changes who gets to be good at this. Not the most technically skilled. Not the most institutionally compliant. The most self-aware, curious and brave. The ones who can articulate how they think and why they make the decisions they make.
“Experienced lawyers have an enormous advantage in this new world, and most of them don’t realize it.”
The Knock-On Effects No One Is Discussing
If Shapiro’s approach scales (and it will) the ripple effects go far beyond efficiency gains for boutique firms.
The mentorship model inverts. Zack described packaging his judgment into a plugin that any associate could install and immediately produce work at his standard. Think about what that means. The traditional apprenticeship model, where it takes years of proximity and correction to transmit a senior lawyer’s way of thinking, compresses into an instruction file. Knowledge transfer doesn’t disappear. It accelerates. The junior lawyer’s job isn’t to slowly absorb the partner’s preferences through osmosis. It’s to exercise judgment on top of a baseline that’s already calibrated to the highest standard.
The value of experience appreciates, not depreciates. This is the most hopeful implication for lawyers fearing replacement, and the one I keep returning to in my teaching. If AI handles execution, then what clients pay for is the judgment that only comes from years of practice. Experienced lawyers aren’t threatened by this technology. They’re the ones with the richest raw material to feed into it. As Zack told me, “Experienced lawyers have an enormous advantage in this new world, and most of them don’t realize it.”
The access-to-justice gap finally has a credible path to closing. Shapiro’s article is about a Yale Law grad, as a two person law firm, punching above its weight. But the same architecture (encoding expert judgment into AI skills) can be pointed at an entirely different problem: the 80% of Americans who face legal issues without any legal help at all. When expert legal reasoning can be captured, verified, and deployed at scale, the question shifts from “can we afford to help everyone?” to “what’s our excuse for not trying?” That’s the work I’ve dedicated LawDroid to, and it’s why this moment matters beyond productivity.
The Transformation Triangle
What Shapiro illuminated brilliantly is the tools layer. But tools alone have never transformed a profession. You need three things working together: Tools (the AI capabilities), Expertise (the human judgment Shapiro rightly champions), and Education (teaching the next generation how to think within this new paradigm and the feedback loop to improve). That’s what I call the Transformation Triangle, and it’s the framework behind everything I do: my building AI, my teaching, my consulting, and the book I’m writing.
Shapiro’s post went viral because he made the tools layer tangible and real. But the reason millions responded isn’t that they want a workflow tutorial. It’s that they want to know: Is there still a place for me?
The answer is yes. A bigger place than before. But only if you’re willing to do the thing Zack did: sit down, articulate how you think, and trust that your judgment is the asset that makes all of this work.
Closing Thoughts
Something is shifting. Yes, technology is rapidly advancing day-by-day. But, what’s really shifting is mindset. Shapiro’s post landed like lightning because the ground was already charged for it. Professionals everywhere are feeling the tremors and looking for a signal that says: you’re not being replaced, you’re being unleashed.
That’s the real story here. Not a workflow. Not a tool. A new relationship between human expertise and machine capability that makes both more valuable. And we’re just at the beginning.
Listen to my full conversation with Zack on the LawDroid Manifesto Podcast, dropping Monday morning. He’s candid, sharp, and genuinely optimistic. I think you’ll hear what I heard: not just a clever practitioner, but someone who stumbled into a truth the whole profession needed to hear at exactly the right moment.
Stay tuned!
Tom Martin is CEO & Founder of LawDroid, Adjunct Professor at Suffolk University Law School, and Author of the forthcoming AI with Purpose: A Strategic Blueprint for Legal Transformation (Globe Law and Business). He is “The AI Law Professor” and writes his eponymous column for the Thomson Reuters Institute.
Related:
Zack Shapiro, The Claude-Native Law Firm (February 27, 2026)
Thomas G. Martin, The Transformation Triangle (August 28, 2025)
Hit reply and tell me — I read every response and it shapes what I write next.
Good interview and guest, thank you for that. I’m also a big fan of Claude, and especially skills. However, I have reservations about using Claude or any of the big models for legal work. Not because of hallucinations (which as Zack mentioned, doesn't have to be a real problem if you supervise it appropriately), but because the terms of service arguably conflict with the attorney duty of confidentiality.
The court's decision in United States v. Heppner recently brought the risk into sharp relief. It contains this pretty scary morsel: “if certain information that Heppner input into Claude was privileged, he waived the privilege by sharing that information with Claude and Anthropic, just as if he had shared it with any other third party”.
My strategy is to build my own legal tools (using Claude Code - of course), and then using those tools to access Claude through the Anthropic API, along with a negotiated zero data retention agreement. Honestly though, you can get pretty far just by making your own tools with AI even if the resulting tools don't USE AI. A lot can be done with deterministic (not AI) automation, when you have your own private developer in Claude Code!
Ideally, self-hostable models will soon be good enough to combine with self-built tools, ensuring work product never leaves your control.
Links below are to the Heppner decision, and Anthropic re ability to seek a zero retention agreement.
https://scholar.google.com/scholar_case?case=14510865396992864515&q=United+States+v.+Heppner,+No.+25%E2%80%91cr%E2%80%91503+(JSR)&hl=en&as_sdt=2006
https://support.claude.com/en/articles/7996866-how-long-do-you-store-my-organization-s-data
Loved the post and thanks for highlighting Shapiro. I'm looking for knowledge and guidance on AI and the law to help guide me as I look for the best ways to utilize these tools.
You might like this one I recently learned of
https://substack.com/@ab1927/note/c-222209383?r=hiq2l