Legal AI is a trap: prepare to have more work

The promise is everywhere: AI will free up several hours a week for lawyers.

But at the same time, 55% of legal departments have flat or declining budgets, 56% consider themselves understaffed, and 46% expect more work to be brought in-house. (Thomson Reuters Legal)

In that context, the hours saved will not turn into free afternoons or teams suddenly available for more strategic work.

They will be filled immediately.

And unless legal departments actively protect this new capacity, AI will simply accelerate the volume of work completed without solving the underlying problems. ❌

You will save time. Demand will increase. ⬆️

You review contracts faster? Sales teams will send you more of them.

You produce a first analysis in minutes? Management will expect more comprehensive answers.

You shorten turnaround times? That new deadline will quickly become the standard.

This is the productivity rebound effect: the faster and more accessible your department becomes, the more people will use it.

You will process more matters without ever feeling less overwhelmed.

Volume will increase. Expectations will increase. Headcount will remain unchanged.

A few months later, you will be just as saturated as before — only operating at a higher speed.

AI does not always eliminate work. It moves it.

An answer generated in thirty seconds still needs to be reviewed, contextualised, corrected, approved and sometimes explained to the business.

Your production work decreases. Your supervision work increases.

That shift can be useful. But accelerating a task does not mean eliminating the workload.

A contract review completed twice as fast creates no lasting capacity if the number of contracts to review doubles at the same time.

The real question is not: “How much time does AI save?”

The real question is:

What are we going to do with that time?

This new capacity can be used in four ways:

Absorb more volume, allowing the company to grow without adding headcount.

Reduce turnaround times, accelerating contracts and business decisions.

Reduce costs, by limiting external spend or the amount of senior time spent on repetitive tasks.

Refocus lawyers on more strategic matters, such as critical negotiations, risk prevention and transformation projects.

But that last outcome will not happen naturally.

You need to decide in advance how the freed-up time will be used. Otherwise, operational requests will immediately consume it.

An hour saved but not explicitly reallocated will always disappear into day-to-day demand.

Most teams have no idea where their gains go :

According to a 2026 Axiom survey of 528 in-house legal leaders, only 7% of teams had deployed AI at scale, while 83% did not know how to measure whether their investment was producing the expected results. (Axiom)

It is therefore easy to say:

We are saving a lot of time with AI.

But how much? On which tasks? Has that time genuinely disappeared, or has it simply moved into review and verification? Has it reduced turnaround times, costs or the number of requests received? Or has it merely enabled the production of more documents?

Without an answer, productivity remains an impression.

By default, you will absorb more volume — usually through low-value requests. Yet this is the least valuable of the four outcomes for your organisation.

Reducing turnaround times cuts inertia and accelerates business.

Reducing costs has a direct impact on the P&L.

Refocusing lawyers on strategic matters can multiply the ROI of the legal department, reduce disputes and better protect the organisation against its most important risks.

🛑 The best automation is the one that prevents the request from arriving

Most AI tools aim to help you respond faster. That is useful, but not enough.

You will only truly scale when certain requests no longer reach you at all.

When a business user can find and reuse a previously approved answer, the workload genuinely disappears.

When an internal rule immediately shows that legal approval is not required, the workload genuinely disappears.

When a previously resolved case can be handled without new research, drafting or arbitration, the workload genuinely disappears.

This is the approach developed by Sandstone in the US and Mirmi AI in Europe: turning precedents and validated decisions into reusable capacity for the business, rather than merely helping lawyers produce work faster.

The goal is not simply to answer the same question 30% faster.

It is to prevent an already resolved question from being asked again.

Instead of absorbing more unnecessary requests, you can maximise the other 3 benefits: faster decisions, lower costs and more time for strategic work.

That is where the ROI becomes highest — and where you finally stop being constantly overwhelmed. ⏳ 🚀

📊 Measure value, not prompts

To determine whether your AI is genuinely improving productivity, track:

• the number of requests resolved without lawyer intervention;
• the full turnaround time between request and decision;
• the percentage of matters escalated to legal;
• the average cost per matter;
• the amount of time genuinely reallocated to strategic work.

The number of users, logins or documents generated measures adoption. Not value.

AI will not automatically make your team more strategic

It will give you additional capacity.

That capacity can be used to absorb more requests and preserve the same dysfunctions at a higher speed.

Or it can be used to eliminate repetitive work and refocus lawyers on decisions that genuinely require their expertise.

Technology will not make that choice for you.

⚠️ The future of legal productivity is not about doing more work in less time.

🧠 It is about deciding which work you should no longer have to do.