Seventy-one percent of solo attorneys are using AI. Roughly a third of them are earning more than they were before.
That gap is the most important number in legal technology right now, and most firms are drawing the wrong conclusion from it. The instinct is to assume the tools are overhyped, or that the benefit will arrive eventually, or that law is simply different. None of those is what the data says.
I put the question to Joshua Lenon, Lawyer in Residence at Clio and the person behind the Legal Trends Report. His answer was more useful than I expected in this week’s episode of Counsel Cast, Why Isn’t AI Making Your Firm Money?
The variable is breadth, not quality
Joshua Lenon described a flip in the data. For years, most firms reported using legal-specific AI. Research tools. Contract analysis. Task-shaped software that does one job extremely well and touches nothing else in the business.
This year, more firms reported using general AI tools than legal-specific ones. That sounds like a downgrade. It isn’t. It means AI has started moving out of the practice of law and into the operation of the firm, and the operation of the firm is where money is made or lost.
The numbers back it up. About one in ten firms report improved revenue from AI. But among firms that describe themselves as widely using AI across their business, 69% report significant positive revenue impact.
Same technology. Different surface area. Completely different result.
So the honest question is not whether your firm uses AI. It’s how much of your firm it actually touches.
AI’s most underrated marketing function is memory
The clearest example Joshua gave had nothing to do with drafting.
Think about a divorce matter with minor children. Ten years later those children approach the age of majority, and that milestone carries consequences. Tax implications. Support obligations. Whatever the separation agreement said about college. Insurance coverage that shifts.
You knew all of that when you closed the file. Then a decade passed.
No reminder system survives ten years reliably. But AI can read the context inside your existing matters, surface the milestone, and bring it to your attention at the moment it becomes relevant. Joshua’s phrase for it was a research librarian raising things over your morning coffee.
From the client’s side, this reads as a lawyer who remembered. From the firm’s side, it’s a marketing message that is consistent, timely, and grounded in something real.
Speed as a competitive position
Joshua told a story about a lawyer who received a prospect call while standing in an airport security line. He said he would take a look and get back to them shortly. He opened the mobile app, pointed it at files already in the system, asked for a strategy proposal based on the firm’s templates, and by the time he reached his gate he had something to send.
Thirty-five minutes from cold call to proposal. The client signed.
I asked the obvious follow-up. If five lawyers all have the same tools and the same documents, what makes any of them different?
Joshua’s answer: it’s not the tool, it’s the person.
Tools converge fast. What does not converge is judgment. Which cases you take. What kind of practice you are building. Whether you know the client’s history. I have clients going into their second decade with me, and when one came back last year I pulled our notes from the original brand build ten years earlier. That conversation was different from the one any other agency could have had, and no software created that advantage.
The pricing problem has an answer already
Every conversation about AI efficiency eventually hits billing. If you bill hourly and the work takes a fifth of the time, you have handed your client a discount they never asked for.
Most firms know this. Very few have changed anything.
Joshua pointed to a Virginia State Bar ethics opinion that says something firms should have been told louder and sooner. Time spent on a task is only one possible way of measuring value. If you charge five thousand dollars for a matter, the fact that it took less time than it used to does not automatically make that fee unreasonable. You just have to be able to explain why it’s reasonable, and expertise, systems, and tool selection are all legitimate parts of that explanation.
The rules already allowed this. Firms have been waiting for permission that was already on the books.
I spent years as a travel agent before this career, in a job where every minute was measured. I was routinely two to three times faster than my colleagues. When I was five minutes late one morning, my manager wanted a word. I asked whether it mattered, given that I was taking three times the calls. Should I slow down?
That’s the hourly model. It punishes your best performer and rewards the slowest one. Every other industry sorted this out. Law is still deciding.
Your pricing is a marketing decision
Joshua shared something from Clio’s current research: there is more price elasticity in legal services than most lawyers assume. Price is not the first question clients ask. They ask how quickly you will get back to them, and whether you have handled this kind of matter before.
But price is a hurdle, and hurdles matter.
When a prospect can find a flat fee or even a clear range before the consultation, the hurdle is gone before the call starts. The wrong-fit clients screen themselves out. The right ones arrive already comfortable. The conversation shifts from negotiating an hourly rate to discussing what they need and what you provide.
Clients want predictability. They are having one of the worst weeks of their lives. Nobody wants an unknown number attached to it.
The AI policy nobody has written
Here is where I think there is real, immediate opportunity.
Fifty-seven percent of solos and 55% of small firms either have no AI policy or don’t know if one exists. Those are the same thing.
Meanwhile, more than 70% of clients say they want to know whether their firm uses AI. Up from the year before.
That’s a gap you can close this month. Joshua has worked with the Virginia and New Jersey state bar associations on this, and the New Jersey Supreme Court has published a model policy. Clio has a template on its blog.
Frame it correctly and it stops being a compliance document. A public AI policy tells clients how you use technology to serve them better. It gives clients guidance on their own AI use, which matters more than firms realize, because clients are already turning to consumer AI tools whether you address it or not.
Writing the policy also forces a conversation inside the firm about which tools are approved and where the limits are. Without that, you get shadow IT. Staff quietly using whatever makes their day easier, with no parameters and no oversight.
And from a marketing standpoint: if half your competitors have nothing, publishing yours makes you one of the few firms in your market that has answered a question your prospects are already thinking about.
Where the human stops
Joshua’s book pick was Co-Intelligence by Ethan Mollick, and the argument fits the whole conversation. AI is not a replacement for people. It’s an extension of them.
We landed on a metaphor I’ll keep using. Terminator or Iron Man. One is fully automated with no human left inside. The other is a person, amplified. Efficiency for its own sake takes you toward the first one, and that is where firms lose the thing clients actually hire them for.
My own line is specific. I outline first. I write the substance. AI comes in afterward to refine, catch what I missed, and check how it will perform in search. The order is not negotiable.
What to do this week
Audit how widely AI actually touches your firm. If it’s one tool doing one job, that’s your answer on revenue.
Pick one matter type and price it on value instead of hours. Document your reasoning.
Publish an AI policy. Use the New Jersey model as a starting point.
Joshua’s closing advice was to lean in and be transparent. That is right, and I’d add one more word. Publish. Transparency your clients can’t find isn’t doing anything for you.