A&O Shearman’s $1.1 billion AI commitment is mostly lawyer time. That is the point.

On November 5, A&O Shearman announced it will commit USD 1.1bn to AI over the next five years. The firm says the majority of that investment will come from lawyer time: each practice will spend at least 3% of its time developing specialist AI tools and services for its area of law.

The firm is not starting from zero. It rolled out generative AI firm-wide in 2022 and built ContractMatrix, a contract tool for in-house counsel, with Microsoft and Harvey in 2023. It says that in September, 96% of its partners and 93% of its lawyers used its AI tools. It is now building a series of agentic tools with Harvey and offering them to corporates, financial institutions and other law firms.

“This significant investment signals where we intend to take the firm over the next decade,” said Hervé Ekué, the firm’s global managing partner.

Read the number carefully

The $1.1 billion is the firm’s own figure, and most of it is time, not cash. Non-Billable asked how much of the headline number will be spent in cash; the firm had not answered by publication.

That makes the comparison with Kirkland & Ellis, which set aside about $500 million in May to build its own AI platform, less direct than the headlines suggest. One figure is largely a budget. The other is largely an allocation of lawyers’ hours.

That difference is not a weakness in the announcement. It is the most useful part of it.

Why lawyer time is the real investment

Software is the part of legal AI a firm can buy. What it cannot buy is the knowledge that makes a tool useful inside a specific practice: which positions the firm accepts in a negotiation, how it structures a filing, which precedents it trusts, where a matter usually goes wrong. That knowledge sits with lawyers, and turning it into a working system takes their time.

A&O Shearman has decided to pay for that time deliberately, practice by practice, and to put a floor under it. Three percent sounds small. Across a firm of its size, it is a large and recurring commitment of the most valuable hours the firm has.

Kirkland made the same bet by a different route: its own build team, with input from 250 of its lawyers. Two of the largest firms in the world have now said the same thing in public. The value in legal AI is not the model. It is the firm’s own knowledge, encoded by the people who hold it.

What it means for firms that are not A&O Shearman

Most firms cannot dedicate 3% of every practice to AI development, and they do not need to. The lesson is not the percentage. It is the structure behind it.

• Choose the workflows that carry the firm’s judgment. Start with work that repeats, and where the firm’s way of doing it is part of why clients hire it.

• Put named lawyer time against them. A tool configured without the lawyers who do the work will reflect the vendor’s defaults, not the firm’s practice.

• Know who owns the result. When lawyers spend time encoding the firm’s knowledge into a system, the firm should know whether that work stays with the firm or with the platform it was built on.

The last point is the one most firms skip. Time spent configuring a platform is still an investment. The question is whether it is an investment the firm keeps.

A&O Shearman will spend five years and a meaningful share of its lawyers’ hours on this. Every other firm faces the same choice at a smaller scale: whose time builds its AI, and who owns what that time produces.

Sources

• A&O Shearman, “A&O Shearman accelerates AI strategy with a USD 1.1bn lawyer-led commitment,” November 5, 2026

• Non-Billable, “A&O Shearman ramps up AI push with five-year plan it says is worth $1.1bn,” November 5, 2026

• Financial Times, “Kirkland & Ellis to spend $500mn building its own AI technology,” May 28, 2026