Legal operations is the most underrated job in the industry

If you want to know which firms and legal teams are going to come out of the next decade ahead, don’t look at the partners and don’t look at the AI. Look at whether they have a serious legal operations function, and whether anyone listens to it. Because legal ops is where the actual transformation either happens or quietly doesn’t - and it’s the most underrated job in the industry by a distance.

I find this slightly funny and mostly frustrating, because the whole conversation about legal AI is happening one level above the people who will determine whether any of it works.

What legal ops actually is

Legal operations is the unglamorous discipline of making the legal function run as a system rather than a heroic improvisation. Process, tooling, data, vendor management, workflow design, the plumbing that connects all of it. In a corporate legal department it’s increasingly a named team. In law firms it’s often scattered, half-owned, or treated as glorified IT support.

And it is precisely the function that the AI era makes decisive, because everything that actually matters about deploying AI in legal work is an operations problem, not a lawyer problem and not a model problem.

Every hard part of legal AI is an ops problem

Walk through where legal AI initiatives actually succeed or fail, and notice that almost every failure point lives in legal operations:

Is the data usable? AI built on messy, ungoverned, inconsistent data fails slowly and expensively. Whether your data is in a state to build on is an operations question, owned by ops, invisible to the partners cheering for AI.

Does it fit the workflow? Whether a tool lands in the real flow of work or sits unused beside it is a workflow-design question. That’s ops. Lawyers can’t design their way out of a badly-integrated tool, and a model can’t either.

Who owns the thing after launch? The maintenance, the feedback loop, the second-year ownership that separates a product from an abandoned project - that’s ops. Without an operations function that owns the thing, the tool ages into uselessness no matter how good it was on day one.

Does adoption actually happen? The training, the change management, the champions, the measurement of whether fee-earners reach for the tool - all ops. The model doesn’t adopt itself and the partners won’t do the unglamorous spreading work. Someone has to, and that someone is operations.

Every single one of those is a make-or-break factor for legal AI, and every single one sits with operations. Which means the function everyone treats as back-office plumbing is, in fact, the function that decides whether the front-office AI investment returns anything at all.

Why it’s underrated

The undervaluing is structural, and it comes from the prestige economy of the profession. Status in law flows to the people who do the lawyering - the rainmakers, the technical experts, the names on the deals. Operations is support, and support is low-status, and low-status functions don’t get the seat, the budget, or the authority their actual importance warrants.

There’s also a measurement problem. Good operations is invisible when it works - the tool just gets adopted, the data is just usable, the workflow just flows. You notice ops only when it’s absent, and “the absence of preventable disasters” is a hard thing to take credit for. So the function that prevents the AI initiative from failing gets no glory when it succeeds, and the partners who championed the shiny tool take the credit that belonged to the ops person who made it actually work.

The firms that get this will win

Here’s my actual prediction. The firms and legal teams that win the AI decade will be the ones that treated legal operations as a strategic function - resourced it, gave it real authority, put it in the room where decisions are made - rather than as the people you call when the system’s down.

Because the gap between firms is not going to be access to AI. Everyone will have access to the same models, the same vendors, the same tools. The gap will be operational capability - the ability to get data in shape, design workflows that fit, own products past launch, and drive real adoption. That capability lives in operations. The firm with a great legal ops function and decent AI will crush the firm with great AI and no operational capability, every time, because the second firm’s great AI will sit unused on the network drive exactly like every previous generation of tools did.

What I’d tell a managing partner

If you run a firm and you’re spending on AI, the highest-leverage thing you can do is probably not buying more AI. It’s hiring, empowering, and listening to a serious operations function - and giving them the authority to say no to the shiny tool that doesn’t fit, and yes to the boring work that makes the right tools land.

The lawyers are necessary. The AI is necessary. But the people who’ll quietly decide whether your firm actually transforms or just buys software are in operations - and right now most firms are paying them too little, ranking them too low, and inviting them into the conversation too late. Fix that, and you’ve fixed the thing most likely to make your AI investment work.