A&O Shearman announced a $1.1 billion commitment to artificial intelligence over five years, planning to fund most of it through billable lawyer hours rather than cash outlays.
How the Commitment Is Structured
The firm expects each practice area to allocate at least 3 percent of its time to building specialized tools and services. With annual revenue near $3.7 billion, that share translates to roughly $110 million a year, or about $555 million across the five-year span.
Because the allocation comes from fee-earning work, the firm effectively redirects revenue that would otherwise be billed to clients. The approach represents a tangible shift from the traditional model where time equals money.
Industry Comparison
In May, Kirkland & Ellis disclosed a plan to invest $500 million in a proprietary platform, funded directly with cash for engineers, data scientists and hardware. The A&O Shearman filing does not separate cash from time-based contributions, making a direct dollar-for-dollar comparison difficult.
Using the firm’s revenue baseline, the 3 percent contribution accounts for about half of the headline figure, while the remainder would depend on future software licensing or other revenue streams that have not yet materialized.
Diverting a slice of productive hours is a concrete cost to the firm. It reduces immediate billable output but signals a longer-term redefinition of what counts as billable work.
Implications for Lawyers
The firm rolled out its generative AI system “Harvey” in 2022 and later introduced a contract-analysis product called ContractMatrix. By September, 96 percent of partners reported using these tools.
“This significant investment signals where we intend to take the firm over the next decade,” said global managing partner Hervé Ekué.
When technology shortens task duration, the savings must be allocated. Clients typically expect lower fees, while equity partners aim to preserve profit levels that currently average $2.9 million per partner. Associates may see their performance metrics evolve.
Fixed-fee revenue has risen 15 percent over the past two years, a gain the firm attributes in part to faster delivery enabled by its tools. Under fixed-fee arrangements, speed improves margins without reducing the invoice amount.
A recent ruling by the U.S. Court of Appeals for the Tenth Circuit upheld a $114,663 fee award, rejecting the argument that AI-generated work was unreasonable. The decision reaffirmed the duty to verify AI output and maintain accurate billing.
Junior lawyers have traditionally built expertise through research, drafting and review. If software assumes those tasks, firms will need new training models and revised metrics for associate contributions.
One question that remains is how the 3 percent allocation will be reflected in individual hour targets and partner compensation. Historically, partnership committees have not counted non-billed development work toward performance metrics.
