A practical guide to choosing the right AI tools for your law firm: Five Insights that surprised us the most
min readWhat lawyers really want from AI (Hint: It’s not just speed)
After 7,000 data points, five clear patterns emerged. Here’s what matters most.
- Legal competence is king: Lawyers assume that AI will be fast and accurate. So what really matters? Our data shows legal competence beats efficiency every time. Once speed and accuracy are baseline, legal-specific features dominate
Think: understanding complexity, using correct terminology, and handling nuanced queries. - Drafting is the differentiator , granularity matters. Drafting was only 7% of prior usage, but it was the strongest predictor of preference.
Why? Because drafting exposes AI’s limits. Templates help, but from-scratch drafting? That’s where gaps show. - Power-user bias is real – Include diverse personas. Your most enthusiastic users aren’t your baseline.
Our data shows power users scored products up to 0.7 points higher across tasks.
Include occasional users, or risk inflated scores. - Role-based priorities matter ie partners vs trainees. Partners want accuracy and depth of legal reasoning. Senior Associates prioritised long-form drafting capability and reliable citations. Associates prioritised drafting quality and depth of explanations. Trainees want clarity of explanations and usable first drafts. One-size-fits-all assessments don’t work. Role-based criteria = better adoption.
- Disputes capability lags – ensure you test for practice-specific needs. Horizontal AI serves most practices, but disputes lawyers scored products lower overall. Depth of reasoning and citation reliability remain weak spots. We’ll share what this means for vendors soon.
The full report can be found here: A practical guide to choosing the right AI tools for your firm.
👉 Next post: What law firms and vendors should do when choosing an AI vendor
