Most of the law firm AI conversation is happening in the wrong register. The question everyone is asking is “will AI replace lawyers?” The question that matters is “which parts of the law firm are actually running on processes that were designed before the internet?”
The answer, at most firms, is almost everything that touches client acquisition and intake.
What we saw at Zuber Lawler
We spent 18 months inside one of the most selective law firms operating in the US — a firm with a strong institutional reputation, a selective client base, and a practice that had never needed to market itself. That last point is important context: this was not a firm with broken systems. It was a firm with systems that had been built for a different era.
What we found when we mapped the front office wasn’t dysfunction. It was archaeology. Intake processes that had never been formalized. Content and SEO that had been handled incidentally. A web presence that reflected where the firm had been, not where it was going.
None of this required AI to fix. Most of it required design.
What AI is actually good for in legal contexts
The AI applications that have worked in professional services engagements we’ve been part of fall into a few categories:
- Research acceleration. Not replacing research, but reducing the time between “here’s the question” and “here’s the relevant authority.”
- Drafting first passes on standardized work. Engagement letters, NDAs, routine contracts — work where the variability is low and the cost of a first pass is real.
- Content generation at scale. Thought leadership, practice area descriptions, SEO content — not ghostwriting the partner’s analysis, but producing the scaffolding that partners then edit.
- Intake qualification. Screening and routing inquiries to the right people faster, with more consistent questions asked upfront.
What AI is not good for, in our experience, is anything that requires the firm’s specific institutional knowledge — the judgment that comes from years in a practice area, the relationships, the read on a particular judge or market.
The pace problem
The most important thing we learned over 18 months at Zuber Lawler is that law firm AI adoption has to move at the pace of the firm’s culture, not the pace of the AI hype cycle.
A firm that has earned its reputation over decades is right to be cautious about systems that could create a client service failure or a professional liability issue. The advisors who come in treating that caution as an obstacle to overcome are the ones who fail. The ones who treat it as design constraint produce work that actually sticks.
The firms that will get the most out of AI over the next five years are not the ones moving fastest. They’re the ones moving most deliberately — rebuilding systems in the right order, with clear ownership, and with a real understanding of what the firm’s culture will and won’t absorb.
That’s slower. It’s also much more valuable.