Will AI really change the legal industry? It is the question in every room, every conference, every board meeting. And it is probably the wrong question.
A better question is: what will AI change about what clients value and what they are willing to pay for?
The efficiency trap
The current conversation about AI in law is heavily focused on efficiency. Work that used to take three weeks now takes three days. Documents that required hours of review can be analysed in minutes. The promise is clear: faster, cheaper, more consistent.
But there is an uncomfortable question buried in that logic. If a matter that used to take three weeks now takes three days, is that worth less, or more?
If a deal closes before a critical deadline, if a regulatory issue is resolved before it becomes a headline, and if advice lands early enough to change a decision rather than just validate one, that is not less work. That is better work.
The issue is not alternative fee models. It is that the profession is still anchoring pricing to effort instead of outcome. And outcome is contextual: how time-critical was the matter? What risk did it remove? What opportunity did it unlock?
Efficiency does not automatically reduce value. Sometimes it concentrates it.
The better question might be: what would it have cost if the firm had been slower?
The tacit knowledge problem
There is a deeper challenge that the AI conversation is only beginning to address. Firms can feed a model everything that has been written down: case files, precedents, contracts, judgments. But the most valuable knowledge in a law firm has never been documented.
Why did that partner choose one expert witness over another? Why did the closing argument pivot in week three? What did the lawyer read in the room that changed their approach?
That knowledge lives inside lawyers' heads. It never gets captured. And right now, it never gets anywhere near the model.
This is not a limitation of legal AI specifically. It is a fundamental challenge of AI in any expertise-driven profession. The technology works brilliantly with structured, documented information. It struggles with the tacit knowledge, the judgment calls, the contextual reasoning, and the why behind the decision that separates good advice from great advice.
The firms that figure out how to extract that context systematically, not just what their lawyers did but why, will build something that genuinely compounds over time. That is the moat: not the model, not the data, but the captured reasoning of their best people.AI for the business of law
For too long, the AI conversation in legal was almost entirely focused on the practice of law. How AI changes the work, the associate pipeline, and billable hours.
That conversation matters. But it is only half of it.
The more interesting shift is AI applied to the business of law. Pricing. Revenue recovery. Client profitability. The decisions firm leadership and business services teams make every day, mostly with incomplete information.
Industry estimates suggest the top 200 firms leave around 20% of available revenue on the table annually. AI can help close that gap. But only if the data underneath is in good enough shape to work with.
And for most firms, it is not. One client sitting under four different codes. Time entries inconsistently described. Matter data in the wrong category. Before anyone can analyse anything useful, they first have to reconcile the underlying data.
The direction of travel is right. The data has to catch up.
The legal tech landscape is shifting
The broader legal technology market is responding to these pressures. For a long time, the line between practice-of-law technology and business-of-law technology was relatively clear. That line is blurring fast now.
We are watching well-funded technology companies push away from practice-of-law tools and towards the business of law, wrapping intelligence around BD and CRM, pricing and matter data, workflow and operations. This is fertile ground. Much of the business-of-law technology stack is still built on legacy systems that are being marketed with AI labels rather than fundamentally re-engineered.
When the underlying AI models become commoditised (and they will) the differentiators will be workflow, data context, and trust. The firms that invest now in getting their data in order and their business processes aligned will be the ones best positioned to benefit.
What this means for firms today
The practical implications are not about choosing an AI vendor or launching a pilot project. They are about three things:
Rethink how you price. If AI makes your firm faster, that should benefit both the client and the firm. Clients expect to see efficiency gains reflected in fees. Firms need to find pricing models that capture the value of speed, certainty, and risk mitigation, not just hours spent.
Get your data in order. AI tools are only as good as the data they work with. Firms that invest in data quality, standardisation, and integration now will be dramatically better positioned to benefit from AI in the business of law over the next three to five years.
Capture what your people know. The most valuable thing a firm can do with AI is not automate what already exists in documents. It is to start capturing the reasoning, judgment, and contextual knowledge that currently walks out the door every time a senior lawyer retires.
At Beyond Billable Hours, our workshops help firms navigate these shifts practically, from martech stack assessment to data strategy to firm-wide growth planning. If your firm is thinking about how AI changes the business, not just the practice, let's talk.