Why I'm Writing About Contracts, Data, and Legal Tech

Introducing Jack of Domains

I’m David Rosen and I’m launching a Substack, which means I’ve officially reached the stage of my career where I think people want to read my thoughts on a regular basis. I’ve never been accused of lacking self-confidence, so this was probably inevitable.

This Substack will focus on where contracts, data, and legal technology meet. You’ll read about why I think contracts are the most structured data in business (despite people calling it unstructured data), why generative AI is both the biggest opportunity and danger in legal tech, and what I learned about building teams as a manager. There will also be war stories, poker stories, bad jokes, and the occasional Boston sports reference. You’ve been warned.

Before I explain why you should care what I have to say, let me tell you how I got here.

I was a database developer before I became a lawyer. That path gave me a perspective that’s been valuable throughout my career. I started at Cleary Gottlieb, then moved in-house to Credit Suisse, UBS, and Bridgewater Associates, where I worked as Senior Counsel. I spent decades in finance, working on problems that required both legal precision and technical understanding. Then, in 2019, I co-founded Catylex, a contract intelligence platform that transforms entire legacy contract portfolios into searchable data.

That background matters because the contract data problem sits at the intersection of three domains: legal, business, and technology. Most people who write about this space come from one of those domains. I’ve spent meaningful time in all three. I’m a good lawyer, but there are many better ones. While I’ve managed crises and people in some deep salt mines, I never made it to the top of anyone’s org chart until I drew my own. And I’m tech-savvy enough to be dangerous, vibe-coding before it was a thing. I once talked IT at a no-longer-living bank into giving me admin privileges (not correlated, I hope). This mix of experiences shapes how I think about contracting problems and the solutions. Hence the name Jack of Domains.

A technical background in law is not unheard of, but it’s fairly uncommon. I didn’t realize what an oddball I was in the legal space for many years. When I got into derivatives work, I immediately saw the parallel: intricate contracting where you’re looking at books of definitions incorporated by reference, where a five-page contract is really a 500-page contract because it pulls in different modules. That’s the same exercise as coding. When you write good code, you write modules, pass through parameters, modify what you need to, and sometimes go focus on the definitions when they’re not good enough. I always thought of what we did as lawyers as actually code, and that was a huge influence in the way I approached the legal field.

I thought of my technical experience as a side show, a road started but not followed. Among other things, I hadn’t been directly involved in a tech project for decades, and didn’t realize how hard it was to implement tech in a real business. The moment it really clicked that it was something else was during a contract database project at Bridgewater.

I inherited the project as a business owner and head of counterparty legal (the lawyers who faced Bridgewater’s trading counterparties). My only responsibility was to make sure the questions we were asking of the tech were the right ones, and the answers fit for purpose. At one point, I looked around the room and realized I was the only person who simultaneously understood what the lawyers were saying, what the technologists were struggling with, and what management was actually asking for.

I ended up seeing something through my older technical lens that didn’t make sense to me and began probing. I initially got pushback, essentially “stay in your own lane”. As the issues snowballed, everyone realized the project was about to crater and lead to a potential mass sorting (translation from Bridgewater would be plank-walking).

I’m happy to say we all avoided that fate. Everyone, including Jamie Wodetzki, my vendor at the time and now CPO at Catylex dug deep and got the project over the line. I also realized the overlap of domains wasn’t just interesting, it was where I should be.

What You Can Expect Here

This Substack will apply a cross-domain perspective to several themes that in my view don't get enough thoughtful discussion:

The gap between what legal tech promises and what it delivers

There's a lot of hype in legal tech, particularly around AI. I'll dig into what delivers, what falls short, and why. I’ll look at the difference between tools that primarily address operational efficiency versus those that provide decision-grade visibility into risk and exposure. I recently took a stab at that on our company blog, exploring a general counsel's perspective on legal technology vs their core responsibilities.

The parallels between contract data and code

Contracts are structured documents, even though they’re written in natural language. They have syntax, conventions, and patterns. Understanding those patterns is essential to extracting meaningful data at scale. Thinking about contracts as structured data changes how we approach contract analytics.

The business case for contract intelligence

This isn’t just a legal problem. The way companies structure themselves (what the CFO tracks, what the chief risk officer tracks, what sales tries to track in the CRM) has one golden source: the contracts. The relationships organizations care about, their constraints, what they agreed to do, what they can expect to happen are buried in their contracts, but it’s so hard to get the data out. I believe firmly that if you can really crack the data problem at the golden source of all that data, it will flow through to all the other things. I’ll look at how different stakeholders (GCs, CFOs, CROs, procurement leaders) think about contract data and what they need from it. I am very interested in the specific issue of translating legal risk into ROI, dollars and sense, to bridge the domain—I’ve got more on this coming, but I was recently inspired to write a LinkedIn post on Sateesh Nori’s application of objective measurement to efficiency and outcomes for legal aid organizations using Claude Code (an ironically non-profit use case that applied metric-based ROI analysis to legal tech).

The implementation challenge

Many organizations struggle with contract data extraction projects. They're expensive, time-consuming, and often fail to deliver the promised value. I'll examine what makes these projects succeed or fail, and how to approach them differently.

The future of legal work

As contract intelligence becomes more accessible, the role of lawyers will evolve, impacting legal departments, law firms, and the profession more broadly. I also have a prediction: I think we’re going to see the death of contracts as we know them. It may be years away, but we’re not going to see contracts that are written in paper. The end state may be organizational preferences that mesh and generate a contract behind the scenes, with lawyers no longer drafting documents.

Generative AI: opportunity and danger

The advent of large language models has changed the conversation around contract intelligence, legal tech, and knowledge work more broadly. I’ll write about what I think generative AI gets right, what it gets wrong, and where the real opportunities and dangers lie. This isn’t just about contracts—it’s about how we think about knowledge work, expertise, and what humans bring to the table that machines don’t—note the pattern of this sentence, which looks like generative AI wrote it; it did not (or did it?)

People management and building teams

I spent what felt like dog years as a manager at Bridgewater, and then as a startup co-founder building teams across functions: legal, engineering, sales, customer success. I’ve learned a lot about what holds up and what doesn’t when it comes to hiring, managing, and building organizations. I’ll share what I’ve learned, what I’ve gotten wrong, and how I think about building teams.

Other things that might come up

Poker stories could happen. I’m 240,091st in the Hendon Mob’s all-time poker money list, which isn’t much of a brag because it only shows winnings, not amounts blown on entries. Bad jokes are likely (apologies in advance). Case in point. I’m a big sports fan, particularly Boston teams, so you might see some of that bleed through. I know people have opinions about that. Bring it on.

Why This Matters Now

I’ve spent the last several years building a company that solves this problem. Along the way I’ve discovered that the conversation around contract intelligence often misses the most important points.

We’re at an inflection point. The technology to disrupt the legal domain is already here; what remains to be seen is the full shape and extent of that disruption, and how we deal with emergent challenges around accuracy, traceability, and trust. Of course the market will ultimately sort out which approaches survive, but many end-states are possible.

This Substack is my attempt to contribute to this conversation in a way that’s thoughtful, honest, and useful. I’m not here to pitch a product or promote a particular approach. I’m here to explore ideas, share what I’ve learned, and engage with others who are thinking about these problems.