AI is moving fast enough that even the people building it (say they) want to slow it down. Waiting for things to settle is tempting, but firms can start preparing now with four steps: decide who leads, set a usable policy, give everyone access, and train your people.
Earlier this month, Dario Amodei, the CEO of Anthropic, published an essay called “We Must Pace the Frontier.” He argues that AI companies should slow how fast their models improve, and he calls on governments to help, including through regulation.
I won’t weigh in here on the debate about AI regulation, risk, and regulatory capture. My point is simply that this technology is developing at a breathtaking pace.
If the people building AI think it’s moving too fast, it makes sense why many law firms are choosing to sit back and see how this plays out. I understand the instinct. The tools change every few months, everyone is busy, and nobody wants to invest in something that will look different next year.
I think waiting is a mistake. The technology will keep changing whether firms wait or not, and the firms that start taking proactive steps now will be ready for whatever comes next. Here are four things every firm can do right now.
Every firm needs a group driving this: a committee that experiments with the tools, and a champions program that brings everyone else along. This is change management, and it needs both support from firm leadership and energy from the middle of the firm.
Next, there’s the question of who owns the transformation. It’s tempting to give decision rights to IT and security, and depending on the people, that may be the right call. But every part of a firm has its own incentives. If your security team’s job is to keep the environment secure at all costs, firm leadership should ask whether that’s the right home for a change that is going to make them uncomfortable. Some security teams make those trade-offs well. If yours can’t, name a chief innovation officer or a chief AI officer, or deputize someone and give them the authority to shake things up.
It will be uncomfortable. If you wait until everyone feels comfortable, the world will pass you by.
If you don’t have an AI policy, your people are using AI. If your policy is too restrictive, your people are using AI. They’re just not telling you.
A good policy tells lawyers which tools are approved, what information can go into them, and who to ask when they aren’t sure. Write it around how your lawyers work, so that following it is easier than working around it.
Give every person in your firm access to an enterprise-level AI tool. When the firm provides the tool, you can supervise how it’s used, monitor it, and protect your data. Some firms leave each lawyer to set up an account on their own, which leaves the firm with no view of which tools are in use or where its information is going.
“Everyone” includes your business professionals. Some of the biggest efficiency gains in any firm come from freeing up non-billable time. Your finance team could be doing higher-value work than it does today, and your billing team could be doing more proactive work.
For instance, I’ve built a workflow that reviews prebills and identifies categories of unbilled time. But if your finance team doesn’t have access to (and training on) AI, your firm can’t implement that solution across the firm.
You’re welcome to discount this one, since training is what I do. But if you don’t teach your people how to use the technology, they won’t know how to use it.
Firm-wide training does two things. First, it smooths the jagged edge of adoption and skill within your firm by giving everyone a shared baseline. Second, firm-wide training gives everyone shared frameworks and mental models for how to use AI. This is what makes collaboration easier. If every lawyer has a different way of prompting and a different approach for structuring context, each one is working from a different mental model of the technology, and your power users end up working in silos.
When everyone starts from the same foundation, a partner and an associate can hand work back and forth without first working out how the other one uses the tool.
None of these four moves requires waiting to see where AI lands. Pick the one your firm hasn’t done yet, and start on it this month.
I write about this every week in The Bell Rock Brief. If you’d like the next one in your inbox, sign up below.