The economics of legal work are changing from the bottom up

When people talk about AI changing the economics of law, they tend to look up - at partners, at billing rates, at the headline price of advice. I’d look the other way. The change that actually matters, the one that reshapes the whole machine, is happening at the bottom of the firm, in the work that juniors do. And it breaks something load-bearing, because the entire economic model of the traditional firm is built on that work being done by people.

The leverage model, briefly

The traditional law firm is a leverage machine, and the leverage is the business model. A partner doesn’t make money mainly from their own hours. They make money from the gap between what associates are paid and what associates are billed. You hire a pyramid of juniors, charge clients more for their time than you pay them, and the spread - multiplied across a lot of bodies - is the profit.

This is why the work juniors do is economically sacred even when it’s intellectually thankless. The document review, the due diligence, the first drafts, the bundle preparation - that high-volume, repeatable work isn’t a cost centre the firm tolerates. It’s the engine. The leverage runs on it.

Now ask what happens to a leverage machine when the thing being leveraged - junior hours on repeatable work - can be done by software at a fraction of the cost and time.

The squeeze nobody wants to name

What happens is a squeeze the firm finds very hard to talk about honestly, because every honest version of it is uncomfortable.

The repeatable junior work is exactly the work AI does first and best. So the firm can now do it with far fewer junior hours. That sounds like pure margin - and short term, it is. But the leverage model didn’t just use junior work to make money. It used junior work to train the next generation of partners. You became a good senior lawyer by doing thousands of hours of the grindy stuff and slowly absorbing judgement through your fingertips.

Pull the repeatable work out from under the juniors and you’ve improved this year’s margin while quietly defunding the firm’s own future. The pipeline that turned juniors into partners ran on the very tasks you just automated. Nobody wants to name this, because the short-term incentive (do more with fewer juniors) is in direct tension with the long-term one (you still need to manufacture senior judgement somehow).

The pricing follows the cost, eventually

The second-order effect is on price, and it arrives more slowly but just as surely. As long as only the firm knows that the due-diligence exercise now takes a tenth of the hours, the firm keeps the difference. That’s the early-mover margin, and it’s real.

But clients are not stupid, and they have their own AI now. Sophisticated buyers - general counsel who’ve productised their own functions - increasingly know roughly what the repeatable work should cost once it’s automated. The information asymmetry that let firms charge judgement prices for assembly work is closing. The price of the bottom-of-the-pyramid work is going to fall toward its new cost, and the leverage profit that depended on that work being expensive falls with it.

So the firm faces a double change from the bottom: fewer billable junior hours to leverage, and falling prices on the junior work that remains. The engine gets smaller and lower-margin at the same time.

What replaces the pyramid

I don’t think this is the end of firms, but I think it’s the end of the pyramid as the default shape. The economics push toward something flatter and stranger: fewer juniors, doing less of the repeatable work and more of the supervising-the-machine work; more value concentrated in genuine judgement, which AI doesn’t replace; and revenue increasingly coming from productised outputs priced as products rather than from leveraged hours priced by the body.

That last shift is the deep one. The firm that thrives doesn’t just defend its leverage margin as it erodes - it builds a new economic engine on top of the productised work. It takes the now-cheap-to-produce assembly and sells it as a product, at product economics, at scale, instead of mourning the leverage it used to extract from the same tasks by hand. The margin moves from the spread between junior cost and junior bill, to the spread between the marginal cost of a productised output and its price. Different engine, same instinct: find the repeatable work and make money from doing it more efficiently than the client could.

The thing to watch

If you want to know which firms will be fine, don’t watch their partner billing rates. Watch what they do about the junior pipeline. The firms that quietly automate the bottom and bank the margin while doing nothing about how they manufacture senior judgement are eating their own future for a better quarter. The firms that automate the bottom and deliberately rebuild how juniors learn - apprenticing them to the judgement rather than the grind - and and turn the automated work into priced products, are building the next version of the machine.

The economics of law are changing from the bottom up. By the time it’s visible at the top, in the partner numbers, the firms that read it early will already have rebuilt the engine - and the ones that mistook a margin bump for good news will be wondering where their next generation of partners went.