Baker McKenzie Cut Up to 1,000 Jobs to AI. Almost None Were Lawyers

Baker McKenzie, one of the largest law firms in the world, just cut up to 1,000 jobs and pointed at AI. The detail that matters, and that most coverage skipped, is that almost none of them were lawyers.

The cuts, between 600 and 1,000 roles, landed on business services. IT. Knowledge management. Marketing. Secretarial. Design. The firm tied the decision to AI integration, and it is being described as the largest AI-attributed workforce reduction the legal industry has seen. The people who lost their jobs were not the ones in court. They were the ones who kept the machine running behind the lawyers.

If you run a professional services firm of any size, this is the story to study, because the structure Baker McKenzie just reshaped is almost certainly the structure you are built on too.

The part everyone gets wrong about AI and jobs

The common assumption is that AI comes for the expert first. The lawyer, the strategist, the consultant, the specialist whose name is on the door. It is not playing out that way, and Baker McKenzie is the clearest example yet of why.

AI reaches the support layer first, because the support layer is where the repeatable, documentable, systemizable work actually lives:

• Document processing and formatting.
• Research assistance and first-pass drafting.
• Knowledge management, tagging, and retrieval.
• Coordination, scheduling, and administrative chasing.

None of that is the expert's judgment. It is the scaffolding around the expert. And scaffolding, by definition, follows rules that a capable model can now follow too.

Why does the support layer go first?

Because it is the most systemizable part of the business. The work is high-volume, it repeats, and the cost of a small error is usually contained internally rather than exposed to a client. Those are exactly the conditions under which automation works well and adoption feels safe. The expert's judgment, by contrast, is low-volume, high-variance, and expensive to get wrong in front of a client. That is the hardest thing for a tool to touch.

So the firm keeps the fee earners and thins the layer that prepares the work for them. On a spreadsheet, it is the obvious move. It is also the move that quietly changes what a career in the firm looks like.

What this means for your firm

Most service firms are structured like a law firm, just smaller. A few people whose judgment clients pay for, surrounded by a larger layer of people who prepare, format, coordinate, and chase. For years that support layer was simply the cost of doing the work. Baker McKenzie just demonstrated, at scale, that this is the layer where the math changes first.

That cuts two ways, and both matter:

• If your value sits in that support layer, the ground is moving under you faster than the headlines about "AI lawyers" or "AI consultants" suggest.
• If you employ that layer, the roles you have treated as fixed overhead for a decade are the first ones a client or competitor will look at when they run the same calculation.

There is a fair caveat. Some of these cuts were probably coming anyway, and AI is a convenient thing to point at. But that almost makes it worse, because it means the support layer is now exposed whether the AI reason is real or just the acceptable one to say out loud.

The exercise worth doing

Map your own team into two columns. In one, the judgment clients actually pay for. In the other, the support that prepares the work. Be honest about which people and which tasks land where. Then ask which column is growing and which is being quietly automated, both inside your firm and inside your clients.

The goal is not to cut your support layer to match Baker McKenzie. It is to see clearly which roles are exposed, to move the people in them toward judgment work where you can, and to make sure your firm's value is concentrated in the column a tool cannot copy.

If you want these breakdowns a week before they become obvious, the JAS newsletter covers one of these shifts every week for people who run and sell services. It is free, and you can subscribe below.