AI Skills for Lawyers: What to Learn Before Your Firm Decides For You

You're a lawyer. You bill by the hour. And you've probably already worked out that AI tools can do in twenty minutes what used to take a trainee three hours. Which raises some uncomfortable questions about your business model.

But you're here, which means you're past the denial phase. Let's get into what to actually learn.

I'm not a lawyer. i was a data scientist who got made redundant and now works as an AI consultant. But I've worked with several law firms on their AI strategy, and I've sat in enough meetings to know that the legal profession is about two years behind finance in AI adoption and moving fast to catch up. The window to get ahead of this is closing.

Also, i should say: the lawyers who are learning AI tools right now are billing more, not less. Because they're doing higher-value work faster. The ones who aren't learning are the ones whose roles are shrinking.

The skills that actually matter

1. AI-assisted legal research. Not just typing a question into ChatGPT and hoping for the best. Knowing how to use purpose-built legal AI tools to search case law, identify relevant precedents, and produce research memos. And critically, knowing how to verify the output. Because AI hallucinating case citations is not a theoretical risk. It's happened. In court. More than once.

2. Contract review and analysis with AI. Using AI tools to review contracts, identify unusual clauses, compare against standard terms, and flag risks. The skill isn't letting AI read the contract for you. It's building a workflow where AI does the first pass and you apply professional judgement to the exceptions. This is where you multiply your capacity without sacrificing quality.

3. AI-powered document drafting. Producing first drafts of legal documents using AI, with the right constraints. This means knowing how to specify jurisdiction, applicable law, and client requirements in your prompts. It means having a systematic review process for AI-generated legal text. And it means knowing when AI drafting is appropriate and when it absolutely isn't.

4. Legal data analytics. Using AI to analyse litigation outcomes, settlement patterns, and case timelines to inform strategy. Bigger firms are already doing this. If you're at a smaller firm, doing this well is a genuine competitive advantage.

5. Confidentiality and AI risk management. Understanding what happens to data you put into AI tools. Knowing which tools are safe for client-confidential information and which aren't. This is both a practical skill and a professional obligation. Get it wrong and you've got an ethics complaint. Get it right and you're the person the firm trusts to lead AI adoption.

Tools to learn first

Harvey or CoCounsel for legal-specific AI. These are purpose-built legal AI tools and they're significantly better than general-purpose AI for legal work. They're trained on legal data, they cite sources, and they understand legal reasoning in a way that ChatGPT doesn't. If your firm has access, learn them properly. If it doesn't, advocate for a trial.

ChatGPT or Claude for general drafting and analysis. For non-confidential work, general AI tools are useful for initial research, drafting client communications, summarising long documents, and preparing for meetings. The key rule: never put client-confidential information into a tool without understanding its data handling. Use enterprise versions with proper data agreements.

Luminance or Kira for contract review. If you do any volume of contract work, these tools will change your life. They can review hundreds of pages in minutes and flag the clauses that need your attention. Learn the customisation features. Set up your own clause libraries. The more you teach the tool, the more useful it gets.

How to demonstrate these skills

Run a time comparison. Take a task you do regularly... research memo, contract review, document drafting. Do it your normal way and time it. Then do it with AI tools and time it. Present the comparison, including quality assessment, to your team. Partners love data. Give them data.

Create AI usage guidelines for your team. Cover what tools are approved, what data can go into them, what review processes are required, and how to handle the inevitable errors. This positions you as the responsible AI adopter, which is exactly the person firms want.

Produce an AI-assisted piece of work and show the process. Not just the output. The process. Here's what i asked the tool. Here's what it produced. Here's what I changed and why. Here's the final product. This is how you prove that AI makes your work better, not lazier.

Catch an AI error and document it. This will happen. When it does, write it up. This proves you're using AI responsibly and that your professional judgement is the quality layer. Partners who are nervous about AI will find this very reassuring.

The 1-hour weekend project

Find a publicly available contract. Anything... a terms of service, a standard commercial agreement, a lease. Paste it into Claude or ChatGPT. Ask it to identify the five most unusual or potentially problematic clauses and explain why.

Then read those clauses yourself. Does the AI's analysis hold up? Did it miss anything? Did it flag things that are actually perfectly standard?

Do this with three or four different contracts and you'll develop a very practical understanding of where AI contract review is strong and where it's weak. That knowledge is worth more than any webinar.

One thing that might surprise you: AI is often quite good at spotting missing clauses. Things that should be there and aren't. That's useful.

What to do on Monday

Talk to your firm's IT or innovation lead about which AI tools have been approved for use. If none have been, that's a problem and also an opportunity. Offer to be part of the evaluation process. The lawyers who shape their firm's AI strategy are the ones who'll benefit most from it.

If you're wondering about the bigger picture, have a read about what's happening to legal roles and check out the lawyer role page. But the most important thing is to start using the tools. Theory without practice is just procrastination in a suit.