A&O Shearman Commits $1.1 Billion and Billable Hours to AI Build-Out

A&O Shearman will spend $1.1 billion (£800 million) on artificial intelligence over five years, and — more strikingly — will divert a slice of its own lawyers' billable hours to build the technology, the firm announced this month. The figure is the largest single AI commitment disclosed to date by a Magic Circle firm. Whether it buys genuine technological advantage, or simply a very expensive experiment, remains unproven.

What has the firm actually decided?

The plan has two components, and the second is the unusual one. First, the money: more than 5% of annual revenue will flow into AI ambitions across the next five years. Second, the labor: lawyers in every practice must set aside at least 3% of their billable time to help develop the technology. Put plainly, a lawyer billing 2,000 hours a year will spend roughly 60 of those hours on AI development rather than client work. Across a firm of thousands, that is a substantial reallocation of fee-earning capacity.

The stated rationale is to place lawyers at the centre of AI development, using their specialist knowledge to create tools that can perform tasks within their own practice areas and, in the firm's words, change how legal services are delivered to clients. A&O Shearman says the approach follows several years of experimenting with the technology, with an emphasis on building AI systems around its lawyers' expertise rather than relying solely on ready-made products. Notably, the firm has not said what portion of the $1.1 billion covers the billable-hour diversion, nor how it will measure whether the tools deliver returns.

Global managing partner Hervé Ekué framed the investment in expansive terms. He said: "This significant investment signals where we intend to take the firm over the next decade. AI lets us bring an enhanced skillset to our clients' work, however complex, on a global scale and in a way that wasn't possible before." He added: "The firm that leads will combine exceptional legal judgment with the speed and scale to apply it consistently, anywhere in the world. We intend to be that firm and that is what this investment accelerates."

That is confident language. It is also language that commits to no specific deliverable, no named product, and no timeline beyond the five-year horizon.

How does this fit the wider legal AI race?

A&O Shearman is not a newcomer to the field. The firm began testing Harvey in 2022 and became one of that platform's earliest major law firm customers, before going on to develop AI-powered products of its own. The new announcement shifts the firm from buying and piloting tools to underwriting their construction at scale.

The competitive context matters. Earlier this year, fellow Magic Circle firm Freshfields partnered with Anthropic, giving nearly 6,000 staff across its global offices access to the company's Claude AI models. That arrangement, by contrast, buys access to an external vendor's models rather than committing internal lawyer time to bespoke development. A&O Shearman's move suggests the AI race in BigLaw is no longer only about obtaining the latest tools, but about how much lawyer time firms will commit to developing their own.

What does this mean for practitioners?

For A&O Shearman lawyers, the immediate consequence is concrete: a mandatory claim on at least 3% of billable time, which partners will presumably need to write off, absorb, or reclassify. For clients, the promise is tools built around genuine legal expertise rather than generic platforms — though the firm has not disclosed pricing, expected savings, or how the investment affects fee structures. For competitors, the announcement raises a strategic question: match the spend, or wait to see whether bespoke tools outperform off-the-shelf models at a fraction of the cost.

The takeaway in plain words: a Magic Circle firm has just made its lawyers part-time AI engineers, and the legal market will spend the next five years finding out whether that was prescience or overspend.

By Priya Raman