The headline number is enormous, but what it’s made of is more interesting.
For decades, BigLaw has enshined the idea that time is money and now A&O Shearman is taking that literally by asking its lawyers to spend some of their billable hours building technology that could eventually make a good many of them unnecessary.
On 5 November the firm announced a $1.1 billion commitment to artificial intelligence over five years, but reading the fine print shows that the firm says most of that sum will come from each practice devoting at least 3 per cent of its time to developing specialist AI tools and services.
Not Quite Apples With Apples
The inevitable comparison is Kirkland & Ellis, which said in May it would put $500 million of revenue into a proprietary AI platform over three to four years. Kirkland’s money is cash: engineers, data scientists, chips, and a smaller pot for equity partners.
A&O Shearman has not said how much of its $1.1 billion is cash.
A&O Shearman turns over roughly $3.7 billion a year. Three per cent of that is about $110 million annually, or around $555 million over five years at flat revenue. That goes a long way to explaining how most of the headline figure can be lawyer time before anyone buys a single software licence.
So is A&O Shearman’s bet bigger than Kirkland’s? In headline dollars, yes. In dollars that leave the building, nobody outside the firm knows.
That doesn’t make it trivial. Diverting 3 per cent of fee-earning time is real revenue forgone, and arguably a more radical statement than writing a cheque. It changes what counts as productive lawyer work.
An Early Mover Doubling Down
This is not a late convert. A&O Shearman says it was the first firm to deploy generative AI enterprise-wide when it rolled out Harvey in 2022. It has since built its own products, including ContractMatrix, and claims 96 per cent of partners used its AI tools in September.
“This significant investment signals where we intend to take the firm over the next decade,” said global managing partner Hervé Ekué.
Who Gets the AI Savings?
When technology cuts the time a job takes, someone pockets the difference. Clients expect lower bills. Partners want to protect profits per equity partner, currently $2.9 million. Associates will wonder what happens to their targets.
The firm’s own numbers hint at the answer. Its fixed-fee income is up 15 per cent in two years, partly thanks to AI, though most fees are still billed by the hour.
Under fixed fees, faster delivery fattens margins without shrinking the invoice. Clients will tolerate that until they discover the work took a fraction of the time it used to.
The Courts Aren’t Rushing to Cut Fees Either
On 30 October the US Court of Appeals for the Tenth Circuit upheld a $114,663 fee award to Brownstein Hyatt Farber Schreck, rejecting an argument that the firm’s AI use made its hours unreasonable.
The challenger had to show AI actually caused excessive billing, and couldn’t.
That is a fee-shifting ruling, with the court reminding counsel that AI output must be verified. Accurate billing remains a professional obligation, whatever the software does.
The Bigger Threat Is to BigLaw’s Training Model
Junior lawyers have traditionally learned through research, document review, drafting and due diligence.
If AI absorbs much of it, firms must rethink how associates build expertise and how their contribution is measured. The lawyer who builds a workflow that saves a thousand hours may be worth more than one who records another hundred.
One question the announcement leaves open though is how that 3 per cent counts in individual hours targets and partner pay. Partnership committees, historically, have not rewarded hours that didn’t bill.
The significance isn’t the eye-catching number. It’s that A&O Shearman is redirecting fee-earning time towards technology that could change how those fees are earned, while still billing most work by the hour.
BigLaw spent decades teaching lawyers that time is money. A&O Shearman has taken that literally: it is funding its AI future largely in billable hours, betting that one day the hours won’t be the product.
