Six months ago, I logged into my GC AI laptop on my first day of work. I was joining remotely from Vancouver, Canada, after practicing as a litigator in Big Law. Naturally, the first thing I did was download Outlook and Teams, the platforms I had lived and breathed as a lawyer. But it was eerily quiet. Where is everybody?
Then I logged into Slack and found out.
In my first week, I had to learn Slack, several other unfamiliar platforms, and a new set of startup norms: work in public channels so others can see and contribute, reserve email for external stakeholders only, and choose the right reaction from a bank of thousands of custom emojis.
Within a month, I had gone from learning Slack to learning how to vibe code. I wrote a blog post about my transition from Big Law to legal AI, using metaphors that included being transported to another planet and running with my team in rocket-powered shoes. At the time, I was visiting GC AI’s San Mateo office for the first time and still wrapping my head around what a pull request was.
It turns out you can go far in six months with rocket-powered shoes alongside an ambitious team. I have moved from learning how legal AI gets built to helping build it. This week, I’m back in San Mateo, working with a team of Applied AI Engineers to level up GC AI’s agent harness. If “pull request” and “agent harness” mean little to you, that’s part of the point.
Here is what those first impressions look like six months later.
1. You Have Agency to Build
My month one reflections focused on the agency to learn. The learning never stops, and that’s by design, thanks to the 1% better leadership principle ingrained in the culture here. What I’ve since realized is that while the learning itself is empowering, learning also empowers you to build, and turn ideas into working product.
When I joined, I expected that my role in product development would involve explaining how lawyers work and providing feedback on what engineers built. Six months in, the collaboration is much more hands-on. I write design documents, test product behavior, diagnose why outputs go wrong, and contribute changes directly to the product. My role is integrated directly with the engineering team, and my legal judgment becomes part of the product itself. Engineers review my technical work, while I bring legal judgment to the features we build. We learn from each other, and the product benefits from both perspectives.
In late June, we shipped US Case Law. I worked on the project from design through testing, and sometimes “took the pen” myself to ship parts of the feature. My contributions included UX changes informed by my litigation experience, aimed at making case law research intuitive — which may sound like an oxymoron if you’re used to traditional legal research platforms. I also helped shape the methodology GC AI follows when researching case law, translating the judgment lawyers apply instinctively into research and analysis guidance the AI agent applies.
Prompts, product specifications, and code are all ways to translate legal judgment into product. That is the agency to build as an R&D Attorney: taking what you know as a lawyer and making it part of how legal AI works.
2. Ambition is Contagious
The billable-hour model means you need to prioritize getting the job done within a reasonable amount of time. Lawyers, who tend to be perfectionists, have to learn the balance of stopping at “good enough” sometimes.
Legal AI has expanded my sense of how high the standard can go when it comes to “good enough”. I’ve noticed that my colleagues at GC AI, from account executives to solutions attorneys, and from engineers to revenue operations, are all constantly striving and stretching. The standard of work here is high, and the ambition is contagious.
The R&D Attorney team recently led training sessions for one another on skills we had developed through our individual projects. We shared approaches to customer interviews, product design, AI evaluation, and even data science. Turns out lawyers can do math! What impressed me the most was how our lawyer brains had stretched in different directions, with each of us finding a different way to apply our legal experience to building product. When one person develops a better way to work, the whole team benefits.
That same energy shows up in day-to-day work. When I share a project proposal, draft deliverable, or even a blog post, I’m often met with some version of: why not 2X it? Six months in, I have started bringing that same question to my own work.
3. Shipping Means Choosing
I’ve always been a yes person, a trait I suspect is common among lawyers. I remember resonating with an idealistic principle from design thinking that I learned in business school: start with "yes” and give ideas room to grow. At GC AI, I have learned the complementary discipline of starting with “no”. An idea has to earn its way into the product.
The engineers and R&D Attorneys gathered in Montreal this summer to plan this quarter’s product roadmap. We brainstormed a long list of features we wanted to ship, then faced the harder question: which ones should we prioritize, and conversely, deprioritize? Prioritization often begins with “no”, or at least “not yet”.
Our customers drove those decisions. The R&D team monitors customer feedback every day, identifies patterns, and carries that signal into product discussions. We combine what customers tell us with our own experience practicing law in exercising our judgment. That allows us to advocate for what busy in-house lawyers need most when priorities are set. Customer obsession means identifying the underlying need and choosing the best way to meet it, whether the answer is yes, no, or not yet.
It’s too bad that “it depends” is less useful in roadmap planning than it is in giving legal advice.
4. The ROI is What You Make It
We value data-driven decisions, and ROI often comes up in our conversations with customers. Lately, I’ve also been thinking about the ROI of working here.
I recently listened to an episode of Lenny’s Podcast in which Max Schoening, head of product at Notion, described why agency matters more than developing any one skill in the AI era. This idea stayed with me in this new chapter of my career and in how I think about being a lawyer. For recent law graduates wondering what AI means for their future, I find it encouraging that learning to use AI meaningfully and developing legal judgment can happen together.
For me, every investment I’ve made in learning a new skill, testing an idea, or taking on responsibility for something unfamiliar has returned greater trust and another opportunity to contribute. At GC AI, initiative has compounding returns. Six months in, I’m contributing in ways I could not have imagined when I joined.
At least once a week, I catch myself thinking “I can’t believe this is what I do for work now”. And that’s not solely because of the empty Outlook inbox I still have.
What else hasn’t changed in the last six months? We’re still building the future of legal AI, and we’re still hiring.
