I asked my AI to draft a contract. It already knew my company, my format, my risk positions, and what to flag without being told. That's not a prompt. Here's what it actually is.

A federal judge just ruled that privilege didn't protect a defendant's conversations with consumer Claude. Most of the coverage told you what he decided. Almost none of it told you what he left open or where the real risk actually lives. It's not the AI-assisted memo. It's what you fed in to produce it. And if you're using Claude Cowork, there's a second problem that has nothing to do with waiver and everything to do with what happens when you have to defend the privilege in court. I went to Anthropic's own support pages to understand the exposure. What I found should be in every in-house legal team's AI policy.

I've had dreams of a live, interactive legal dashboard forever. This week I started building it. I'm also covering how I restructured our Google Drive using Claude Cowork and why I never rely on just one AI tool for legal work, including the prompting technique that forces AI to argue against itself.

AI doesn't know your jurisdiction, your risk tolerance, or your edge cases. A context file fixes that. Here's what's in mine.

The team was already debating how to respond before I opened Slack. Wordsmith had reviewed the contract overnight. By the time I got there, the business had the information they needed and had already decided to push back. They didn't need me.

That contract showed me something I hadn't fully articulated yet. Every request that comes into legal falls into one of three buckets. Most teams are treating all three the same way.

I woke up to find Wordsmith had reviewed a contract overnight and flagged it as one-sided and way outside industry standards. The team had already read it. They were already thinking through the implications. Without me. That was the moment I understood what a neutral tool actually does.

Legal works in the gray. When legal analysis IS binary — is this a contract or not, is this clause standard or not — we're usually not the ones spending time on it. We spend time on the areas that are genuinely contested. The judgment calls. Which is exactly where AI is hardest to verify.

I'm a top-down thinker — I start with the answer before I fully understand the question. Turns out that's an advantage when you're building with AI. This is my actual legal intake playbook: how I handle problems when they hit Slack, how I decide between making a call and fixing a process, and the three-step framework I use to get AI doing the heavy lifting on broken workflows.

Field Note #2. This week: I stopped trying to stay organized and built a knowledge system instead. How a simple two-output framework, an AI-powered Slack channel, and a shift in how I think about legal documentation is changing how my legal function operates and why most teams are stuck answering the same questions on repeat.

Bonus Field Note: The number one thing holding lawyers back from AI isn't the technology. It's not knowing where to start. This week: how building a video game for my six-year-old taught me more about AI workflows than most legal use cases have — and why low-stakes experimentation is the fastest path to high-stakes confidence.

Field Note #1. Three weeks back as GC and I'd already caught myself doing something I tell other lawyers not to do — using AI to confirm what I already thought.

This week: flipping your prompting perspective, why I'm not replacing our classification tool with gen AI, and what happens when you can query your own codebase. The stuff nobody's talking about.

I'm a General Counsel building an AI-first legal function at a scaling tech company — and I'm documenting the whole thing in public. Not theory. Not predictions. The actual work: what I'm building, what's breaking, and what I got wrong. Every Friday, free.