How legal services will actually be delivered in 2030

Predictions about the future of legal services tend to come in two unhelpful flavours. The science-fiction version, where AI lawyers argue cases and the profession evaporates. And the cynical version, where it’s all hype, regulation saves everyone, and nothing really changes. Both are comforting in their own way - one because it’s too far off to worry about, the other because it lets you do nothing - and both are wrong.

Here’s my attempt at the concrete, un-dramatic middle: how legal work will actually be delivered in 2030, given what’s genuinely happening rather than what makes a good headline.

Most legal work will be delivered as a product

The biggest change by 2030 won’t be dramatic AI. It’ll be a quiet shift in the form legal services take. A large and growing share of legal work - the high-volume, repeatable categories - will be delivered as products rather than as bespoke services. You won’t engage a lawyer to do the routine thing; you’ll use a product that does the routine thing, with a lawyer accountable behind it.

This is already starting in narrow categories, and by 2030 it’ll be the default for a meaningful slice of legal work: the standard contract, the routine filing, the common compliance check, the predictable transaction. Delivered as a product, priced as a product, available at a speed and cost that bespoke service can’t touch. The lawyer doesn’t disappear - they move behind the product, supervising and standing behind it, handling the exceptions. But the unit of delivery changes from “hours of a person” to “use of a product,” and that single change reshapes everything around it.

The bespoke work survives, and gets more valuable

The genuinely bespoke, high-judgement work - the bet-the-company litigation, the novel transaction, the situation no precedent fits - doesn’t get productised, and in fact gets more valuable, not less. When the routine work is commoditised into cheap products, the scarce, high-judgement work is what’s left to differentiate on, and clients will pay accordingly.

So 2030 looks barbelled. At one end, productised legal work at product economics - cheap, fast, scaled. At the other, premium human judgement at premium prices. What gets squeezed is the middle: the work that’s too routine to justify premium human prices but was historically delivered that way because there was no alternative. By 2030 there’s an alternative, and the middle - which is where a lot of current firm revenue quietly lives - gets hollowed out. The barbell isn’t a forecast so much as a description of where the pressure is already pointing.

The delivery model: supervised systems, not autonomous AI

How does the productised work actually get delivered? Not by autonomous AI making legal decisions unsupervised - regulation and liability won’t allow it, and they shouldn’t. It gets delivered by supervised systems: AI doing the assembly and the application across many steps, with humans accountable, checking, and handling exceptions.

The 2030 fee-earner, on the productised work, spends much less time doing the work and much more time supervising the system that does it and dealing with the cases the system flags. Their value shifts from production to judgement and exception-handling. The good ones are the ones who can quickly tell when the system’s output is wrong and take responsibility for the call - a genuinely different skill from doing the task by hand, and one the profession will have to learn to teach deliberately, because the old apprenticeship that taught it is exactly what’s being automated away.

Who delivers it: a messier ecosystem

By 2030 the “who” is more varied than today. Traditional firms still exist, and the best of them have built productised offerings alongside their bespoke practices - they’ve grown the AI-native business inside themselves. But they’re no longer the only game. There are AI-native legal businesses that don’t look like firms, owning specific high-volume categories outright. There are corporate legal departments that have productised their own routine work and no longer send it out at all. There are products sold directly to businesses for the legal work they used to buy as a service.

The clean picture of “law firms deliver legal services” gives way to a messier ecosystem where legal outcomes are delivered by a mix of firms, products, in-house systems, and businesses that don’t fit the old categories. The work still gets done. It just doesn’t all flow through the channel it used to, and a lot of it never touches a traditional firm at all.

What this means for the people in it

If you’re a lawyer reading this and wondering what it means for you, the honest answer is: it depends entirely on which end of the barbell you’re on. If your value is doing routine work that’s about to be productised, the next five years are a real threat, and the move is to get onto the judgement end or to be the one building the products - not to hope the productisation skips your category, because it won’t.

If your value is genuine judgement on hard problems, the next five years are good for you - the productisation of everything around you makes your scarce skill more valuable, not less. And if you’re someone who can build or supervise the systems that deliver the productised work, you’re in perhaps the best position of all, because that’s the role the whole model is reorganising around and there aren’t nearly enough people who can do it well.

None of this is the science-fiction version, and none of it is the nothing-changes version. It’s the concrete middle: most routine legal work delivered as supervised products, bespoke judgement more valuable than ever, the middle hollowed out, and a messier set of people delivering it. It’s not dramatic. It’s just a different industry from the one we have now - and 2030 is close enough that the firms which will thrive in it are already, quietly, building toward it.