Throughout almost every conversation I have with law firm leadership at the moment, the same question arises – will AI replace junior associates. I think it’s the wrong question to be asking. The firms that get more value out of this shift will be the ones asking a different one, not whether junior lawyers are still needed, but how their role should change.
Adoption has moved fast, and a number of firms we work with have built proprietary systems in-house. Many more have adopted commercial platforms, with Harvey being the clear leader among both AmLaw100 firms and thousands of smaller practices. Nobody is debating whether this happens anymore. The question is now what this means for how firms hire, train and deploy their most junior lawyers.
The work disappears – the role doesn’t
A sizeable share of first and second year associate work sits squarely in the path of automation. Document review, due diligence, elements of discovery, and all high volume, pattern-based tasks are handled well by large language models, and in my conversations with firms, the platforms are faster and more consistent than a team of associates working through the night.
The temptation is to draw a straight line from that fact to a shrinking associate class. If the tasks that used to occupy the first two years of practice can now be done by a machine, why hire as many people to do them.
The reasoning fails because it treats junior work as an end in itself rather than as training. Document review and due diligence are not simply tasks to complete – they are how junior lawyers learn to read a contract properly, spot the issue that matters among a hundred that do not, and build the judgment that eventually makes them useful in a negotiation or a courtroom. There is no shortcut to that education, and every senior lawyer I place laterally has come up through exactly this route.
A partner cannot run a complex financing or lead a contentious deal without having spent years learning to see the detail beneath the surface of a transaction. Exceptional mid-level and senior attorneys are built on rigorous junior level training.
Training becomes an investment, not a byproduct
What changes is the economics of that training, not its necessity. For decades, associate development has just been a natural byproduct of billing hours. Associates learned by doing chargeable work, and firms were, in effect, paid by clients to train their own workforce.
As AI absorbs more of that chargeable, repetitive work, training separates from revenue. It becomes something firms choose to fund on purpose, because they know it produces excellent lawyers a decade later.
That shift carries a cost. Firms that treat it as an easy saving rather than a strategic choice will regret it within a decade. However, this process importantly creates room – freed from high volume, low judgment work, junior lawyers can spend more of their early years on what actually separates a good lawyer from a great one, framing an argument well, reading a room in a negotiation, and working out what a client is really worried about rather than what’s written in the brief.
Firms that use this shift to build judgment earlier, rather than simply cut headcount, will end up with stronger mid-level and senior cohorts than the traditional model ever produced.
Productivity points toward more hiring, not less
AI is absorbing routine work at the same time lawyers are spending more of their time on judgment, negotiation, influence and argument. Together, those two things add up to a marked rise in productivity per lawyer, and that’s a very different proposition from a shrinking workforce. This is the point I keep coming back to with clients.
Firms rarely respond to higher output per employee by hiring fewer people. In my experience, it tends to do the opposite, expanding the volume and complexity of work a firm can take on with confidence. Tomorrow’s junior associate may spend less time doing basic work and more time in the room where the deal actually gets negotiated, but the commercial case for hiring that associate doesn’t weaken. A firm that gets more out of each lawyer has a stronger case for growing its associate base, not a weaker one.
The premium on experience rises
Where I’m seeing the sharper change is in lateral hiring. As AI takes on more of the groundwork, the value a firm places on the judgment, relationships and dealmaking instincts of its most experienced lawyers goes up with it. Senior associates, counsel and partners who can exercise sound judgment under pressure, manage a negotiation to a close, and bring in or hold onto significant client relationships become harder to replace, not easier. I expect demand for that kind of proven talent to keep growing over the coming months, and firms to compete harder for it.
That points to a market splitting into two speeds. Entry level hiring should hold broadly stable, anchored by the ongoing need to train the next generation of senior talent. Lateral hiring at the experienced end will get more competitive, as firms work out that the lawyers who can direct a blend of junior talent and AI agents toward a successful outcome are now doing work of real, lasting value.
A new definition of the standout associate
Using AI tools well won’t be what sets the best associates apart. What will set them apart is the ability to run a deal or a matter as an orchestrator, managing a team of junior associates, mid-level associates and AI agents, and getting the best out of each. That’s a different skill from the one the profession has traditionally rewarded early in a career, and I think the firms that spot and develop it, both in how they train juniors and how they recruit laterally, will be better placed over the coming months than those still stuck on whether AI replaces the associate.
The firms winning the best senior talent a year from now won’t be the ones with the flashiest AI deployment. They’ll be the ones that worked out early that an associate who spent two years learning the fundamentals with a machine doing the heavy lifting is a different, more valuable hire than one who spent two years buried in a data room. That’s the bet worth making now.
