Big Law may be buying AI, but the mid-market may be the one that makes it work
Sarah Saleemi, manager, solution engineering, at Morae explores why SME law firms should carve out their own AI transformation paths, rather than waiting for bigger firms to lead the way
There is a familiar refrain in the mid-market legal sector: we are waiting to see what the big law firms do.
It is usually presented as prudence. Let the global firms buy the licences, absorb the false starts, run the pilots and discover which tools actually work. Then the rest of the market can move with more confidence.
It sounds sensible.
It is also, I think, the wrong lesson to take from what is happening at the top of the market.
Because much of what currently passes for AI leadership in Big Law looks less like transformation than accumulation. The largest firms are plainly spending heavily on technology. But having AI in the infrastructure is not the same as having AI in the operating model. A firm can have specialist legal tools in the estate and still have changed very little about how legal work is actually produced.
Big Law can afford that ambiguity for longer than everyone else can. Large firms have the capital to buy broadly, test in parallel and tolerate overlap. They can keep several tools in play at once, let practice groups experiment unevenly and absorb patchy usability while the market matures.
The mid-market does not have that luxury.
And that may turn out to be its advantage.
For years, the legal sector assumed that scale naturally produced strategic clarity. In this cycle, scale may produce something else: technology sprawl. The larger the institution, the easier it becomes to postpone the harder questions: which use cases genuinely matter, where value is being created, what clients will pay for, and which parts of legal work should actually change.
That is where the firms in the middle have an opening.
Mid-market law firms do not need to win an AI arms race. They do not need the broadest stack or the most impressive vendor list. What they need is a tighter connection between technology and business outcome: fewer tools, narrower use cases and harder measures of success.
The most meaningful gains in legal AI are unlikely to come from having everything everywhere. They will come from applying AI in a handful of places where it materially changes delivery: first-pass review, due diligence, drafting support, knowledge retrieval, matter triage and portfolio reporting. Those are the use cases that alter cost, speed, consistency and client experience.
This is where the mid-market can lead by example.
Not by outspending the global firms, but by showing what practical adoption actually looks like. Mid-sized firms are often better placed to identify high-friction workstreams, move faster on implementation and decide more quickly where AI is creating measurable value. They can use the technology to accelerate repetitive work, shorten turnaround and reduce manual effort. More importantly, they can use that efficiency to unlock better ways of producing legal work for clients: more structured outputs, more predictable delivery, clearer portfolio visibility and stronger business foresight drawn from legal data.
This is the mistake in waiting for Big Law to ‘show the way’. The largest firms may not produce a clean signal for some time, because they do not need to. They can remain in experimentation mode longer and live with tools that are strategically interesting but operationally underused.
A mid-market firm cannot. It has to decide what pays.
That constraint is not always a weakness. Often it is clarifying. A firm that cannot buy everything has to choose what matters and tie AI to margin, turnaround and client value.
The smarter mid-market strategy is not to ask, in the abstract, whether to build or buy. It is to buy selectively and implement relentlessly. Buy where the market has productised useful capability. Then do the harder work internally: redesign workflows, tighten playbooks, define quality thresholds, assign partner accountability and connect the new delivery model to pricing.
In other words, buy the capability and build the discipline.
That is why the mid-market may reach AI transformation before many firms above it: not because it has more money, or even better technology, but because it has less room for theatre.
The next winners in legal services may not be the firms that bought the most. They may be the firms that bought less, implemented better and moved earlier — using AI not as a badge of innovation, but as a tool to redesign delivery, surface sharper business insight and create a more commercially intelligent service for clients.
If that is right, the biggest mistake the mid-market can make is waiting for Big Law to validate the market.
By the time the largest firms have rationalised their AI estates, the more valuable prize may already have gone elsewhere: to firms disciplined enough to turn AI from infrastructure into economics.
