‘Build Your Advantage’: What eBrevia CTO Jake Mundt Thinks Legal AI Gets Right. And Wrong.

eBrevia Co-founder and CTO Jake Mundt speaks at Legal Innovators California 2026.

When eBrevia Co-founder and CTO Jake Mundt speaks, he doesn’t declare that every new model changes everything, and he doesn't speculate about robots replacing lawyers. Instead, he talks about engineering, product development and the very human reality of getting technology to work inside organizations.

That perspective was on display during a panel at Legal Innovators California 2026, where legal leaders from Intel, Chime, Miro and eBrevia compared notes on what AI is actually changing inside legal departments. While the conversation covered everything from adoption to security to the future of legal work, Jake kept returning to a simple idea: AI isn't valuable because it's impressive. It's valuable when it solves a real problem.

Sometimes that means removing work that nobody wants to do.

Discussing the endless cycle of vendor security questionnaires, he laughed that, from eBrevia's perspective, the company's AI is often talking to another company's AI. "Our AIs are just talking to each other at this point," he joked.

It's a funny image, but it also captures where legal AI is creating value today. Those questionnaires aren't difficult legal work. They're repetitive, time-consuming and, as Jake pointed out, they often sit untouched because people dread dealing with them. Automating them doesn't just save time; it helps deals keep moving.

But he's also careful not to oversell what today's technology can do: "Right now it's like automating the pain. But I'm really looking forward to when it can automate value," he said.

That distinction came up several times during the discussion.

Today's AI is excellent at accelerating routine work. The bigger opportunity, in Jake's view, is helping legal professionals make better decisions by bringing together information they would never have found on their own. Imagine software that remembers a counterparty accepted a more favorable clause last year, recognizes that negotiating history in real time and suggests using it in today's conversation. That's not replacing legal judgment. It's strengthening it.

That same pragmatism shapes how Jake thinks about building AI in the first place.

As organizations rush to launch internal AI initiatives, there's often an expectation that the first version should simply work. Engineers know that's rarely how software is built, and Jake believes legal teams should think about AI the same way.

The first version will expose problems. Workflows will need refining. People will discover edge cases nobody anticipated. That's not failure, it's product development.

He put it bluntly during the panel: once you build these systems, you're effectively becoming a software team. They need to be maintained, improved and adapted as your business changes.

That mindset also explains one of Jake's key observations about AI failures.

"Most of the time it fails not because it's not smart enough, but because it didn't have the data you asked about," he noted.

It's an engineer's way of looking at the problem.

Instead of treating every disappointing result as proof that AI isn't ready, ask why it failed. Was the underlying data incomplete? Was the workflow designed well? Was the information stored somewhere the system couldn't reach?

Those are solvable problems, but only if organizations understand where the breakdown occurred.

That willingness to talk honestly about limitations is part of what makes Jake's perspective refreshing. He believes AI is transformative, but he also believes that unrealistic expectations are one of the fastest ways to undermine adoption.

In fact, when the conversation turned to encouraging employees to embrace AI, his advice had very little to do with technology.

"Showing value [means]: 'I solved this thing that annoyed me. Here's how I did it. You can have it too,’" he shared.

People don't adopt AI because someone tells them they should. They adopt it because a colleague demonstrates something genuinely useful. During the discussion, the panelists traded examples of AI show-and-tell sessions, "AI buddies" and even internal leaderboards that rewarded experimentation. Jake's takeaway was simple: practical wins spread faster than mandates.

The same thinking carries over to one of the biggest strategic questions legal organizations are asking today: Should we buy AI or build it ourselves?

Jake's answer surprised more than a few people in the room because it wasn't the one you'd necessarily expect from the CTO of an AI company: "Buy the commodity... build your advantage."

In other words, don't spend time recreating technology that already exists. Buy the infrastructure, the models and the components that everyone needs. Invest your own time building the workflows, institutional knowledge and business processes that actually differentiate your organization.

That's a philosophy that reflects how software has always been built. The best engineering teams don't reinvent every layer of the stack. They focus their energy where it creates something unique.

The conversation closed with predictions about where AI is headed over the next few years. Some panelists talked about disruption, new business models and changing career paths. Jake answered with an analogy.

For years, headlines about self-driving cars focused on the accidents. Now, as autonomous driving improves, the conversation is beginning to shift toward comparative safety. The benchmark isn't perfection, it's whether the technology performs better than people on average.

He believes legal AI will follow a similar path. "That's my long term... 'Oh, you didn't use AI — that's why there's mistakes in your brief.' That's why the person who does use it gets better terms on average when they negotiate," he said.

It's a characteristically understated prediction. Not that AI will replace lawyers. Not that every task will become autonomous. Simply that, over time, using AI well will become part of producing consistently better legal work.

That's also what made Jake's perspective stand out throughout the discussion. While it's easy to get caught up in the latest model release or the newest buzzword, his focus never drifted very far from the fundamentals: solve real problems, understand how technology actually works, keep improving it, and don't confuse hype with progress.

For legal teams trying to separate lasting change from the noise, that's advice worth listening to.

About eBrevia

Established in 2011 and trusted by some of the world’s most prestigious companies, eBrevia is a leader in AI contract analysis, review, negotiation and management with clients in the US, EMEA, and APAC. For more than a decade, eBrevia has served law firms, corporations, audit/consulting companies, and financial institutions, such as Baker McKenzie, Norton Rose Fulbright, Kroll, SAP, Intel, PwC, EY, and MUFG.