What a Lawyer-Turned-Founder Learned About Building in the AI Era
A recap of our recent webinar with Cecilia Ziniti, founder of GC.AI
Cecilia founded GC.AI, a legal AI platform built specifically for in-house counsel. Not for law firms or for general-purpose knowledge work. But strictly for in-house counsel. That specificity is deliberate, and it's the thread that runs through everything she shared in our webinar this week. If you missed it, here are the key insights.
The Insight That Started It All
Before ChatGPT existed as a household name, Cecilia was general counsel at Replit. Working at the intersection of developer tools and the earliest LLMs, she had what she describes as a "jaw-dropping moment" when she asked an early GPT model a legal question and realized the answer was genuinely good. Not "good for a computer." Just good, period.
Getting insider access to powerful AI before most people knew it existed, combined with deep in-house legal experience, gave her a unique vantage point. Her advice to founders mirrors this: the best startup ideas don't come from a whiteboard session. They come from the specific, non-transferable combination of things you've lived through. "You don't need special preparation," she said. "Your existing skillset from your current role already translates."
She also made a point worth remembering if you're still sitting on a startup idea: "Getting promoted at a big company like Amazon is often harder than being a founder."
Why "In-House" Is a Completely Different Market
GC.AI has over 1,500 customers like Logitech, Honeywell, Duolingo, and Liquid Death, and an NPS of 74. That number doesn't happen by accident, and Cecilia's explanation of why gets at something important for anyone thinking about vertical AI.
Law firms are not going to adopt AI aggressively. The reason is structural: their business model runs on billable hours, and AI compresses billable hours. The incentive to automate yourself into efficiency simply isn't there.
In-house legal teams have the opposite incentive. They're cost centers trying to move faster, close deals more quickly, and handle more work without adding headcount. AI is a direct lever on all of that. Cecilia saw this gap clearly and built into it.
The broader lesson: when you're evaluating a market for AI, don't just ask "who needs this?" Ask "who is incentivized to adopt this?" Those are often very different groups.
Legal as a Business Partner, Not a Gatekeeper
The philosophy behind GC.AI isn't just product strategy, it's a deeply held belief about how legal should function inside a company. Cecilia calls it simple: use "we" not "they."
When she was working on Alexa at Amazon and privacy concerns emerged around always-on microphones, she didn't frame it as "legal says we can't do this." She framed it as "we're launching Alexa, and here's how we navigate this together." The difference sounds subtle. The organizational impact is enormous.
Legal teams that embed themselves in product thinking understand the business objective and help find a path to “yes” become strategic assets. Legal teams that wait for permission slips and respond with lists of concerns become friction. AI is actually accelerating this evolution, because the rote document review and research work is increasingly handled by tools. What remains for human lawyers is judgment, relationship, and strategy.
"Vibe Lawyering" and the Democratization of Legal Work
Cecilia coined a term in the session that's going to stick: vibe lawyering, the phenomenon of non-lawyers using AI to handle legal work they never could before. Everyone with access to a good AI tool now has something resembling legal counsel available on demand.
Her framing here is optimistic rather than apocalyptic, and it's backed by an analogy worth remembering: Excel didn't eliminate accountants. It raised the floor on what every businessperson could do with numbers, which actually raised the bar for what accountants needed to bring to the table. AI is doing the same thing to law.
The lawyers who thrive won't be the ones who resisted AI. They'll be the ones who figured out how to operate at a level that AI can't reach: taste, judgment, cultural context, trust. Her advice to law students was pointed: "Being above the API is essential. But the people skills are what AI cannot replicate."
On Hiring, Headcount, and What GC.AI Is Actually Replacing
One of the most honest parts of the conversation was about what actually happens to legal teams when they adopt GC.AI. Hiring slows, but existing counsel takes on more, such as broader scope, faster output, and work that used to require expensive outside counsel overflow.
This is the version of "AI replacing jobs" that doesn't get covered enough: it's less about elimination and more about reallocation. The lawyers who adapt become more valuable. The companies that adopt early close deals faster and need less external support.
Where GC.AI Is Heading
The current product gives lawyers powerful tools for individual tasks like contract review, legal research, and drafting. The next phase is more ambitious with full workflow orchestration. AI that manages a contract from initial draft through negotiation, redlining, execution, and database entry without a human touching every step.
Cecilia drew the analogy to coding tools like the shift from GitHub Copilot (code completion) to Cursor (agents managing multiple contexts simultaneously). GC.AI is building toward the equivalent for legal, which is not just a smarter assistant, but an agentic system that handles end-to-end processes.
The proactive features she described are particularly interesting for anyone building in AI. Imagine your legal AI monitoring your company's existing contracts and surfacing when an FTC guidance update has implications for your marketing team. That's AI as an early warning system, not just a responder.
The Human Stuff Still Matters (Maybe More Than Ever)
The last thing Cecilia shared before we wrapped was a small but telling detail about her company's own growth. A Series B investor committed largely based on a homepage demo video. It wasn’t a polished, produced asset, but an authentic recording with a verbal slip or two that made it feel real.
In a world where AI can generate polished content infinitely, authenticity has become a scarce resource. She uses her real background on video calls. She shows up as herself. And her observation is that audiences of investors, customers, and employees are responding positively.
For founders building in the AI era, that's an important tip. The product differentiation might come from your model or your data or your workflow design. But the relationship differentiation will come from you.
Enrich hosts regular conversations with founders and operators building at the edge of AI. See here for more events. If you found this useful, forward it to someone who'd appreciate it.
