For years, much of the conversation around AI in BigLaw has focused on what the technology means for lawyers. Will it replace junior work? Which tasks will become automated? What skills will associates need to develop?
Now, some of New York’s biggest legal clients are asking a different question: if AI allows law firms to complete work faster, who should benefit from the efficiency?
Goldman Sachs, Morgan Stanley and Citi are reportedly pushing major law firms to rethink their fees as AI reduces the time required for certain types of legal work. Morgan Stanley and Citi are exploring alternative payment arrangements, while Goldman Sachs has reportedly asked firms to quantify the efficiencies they are achieving through AI, with the expectation that clients should share in the benefit.
For BigLaw, this goes well beyond another conversation about legal technology. It raises questions about the economics on which the industry has traditionally been built and, ultimately, the lawyers firms will need as that model evolves.
The Billable Hour Meets AI
The traditional BigLaw model is relatively straightforward. Large teams of associates work on complex matters, their time is billed to the client by the hour, and the difference between what those lawyers cost the firm and what clients pay for their time contributes to the profitability of the partnership.
That model has supported some extraordinary economics. It has also helped create a talent market in which leading New York firms compete aggressively for the strongest associates and partners.
Associate compensation is one example. In 2026, Milbank increased salaries across its associate ranks, with first-year pay reaching $235,000 and senior associates earning well over $400,000. At the same time, average associate billing rates at the largest US firms have risen sharply in recent years.
For as long as demand remains strong and clients are willing to pay those rates, the model works.
AI complicates the equation.
Legal research, document review, contract analysis and aspects of due diligence are among the tasks where new technology can reduce the amount of time required. If work that once generated a significant number of associate hours can now be completed much more efficiently, clients will inevitably start asking whether they should continue paying as though nothing has changed.
That conversation now appears to be moving from theory to the negotiating table.
Wall Street Is Starting to Push Back
The significance of major financial institutions raising the issue should not be underestimated.
Banks are among the legal industry’s most valuable clients, generating work across M&A, finance, regulatory, litigation, investigations and numerous other practice areas. When organisations of that scale begin questioning how legal services are priced, law firms have little choice but to pay attention.
Morgan Stanley intends to move most of its external legal work towards competitive bidding and alternative arrangements such as fixed fees by the end of 2026. Citi is also working towards different pricing arrangements, with its legal leadership arguing that when AI reduces the number of hours required on a transaction, the cost should fall accordingly. (
This doesn’t necessarily mean the end of hourly billing. Complex legal work can be difficult to price with certainty, particularly when the scope of a matter changes rapidly or the value being provided lies in judgement rather than completing a predictable task.
A more likely outcome is a greater mix of pricing structures: fixed or alternative fees for work that can be scoped relatively accurately, with hourly billing retained where complexity and uncertainty make it appropriate.
Even that, however, could have significant consequences for the way firms operate.
What Happens to the Traditional Associate Model?
BigLaw has traditionally operated as a pyramid. Large numbers of junior lawyers perform much of the work required to execute major transactions and disputes, while progressively smaller numbers advance through the senior associate and partnership ranks.
If AI removes or dramatically reduces some of the work historically performed by junior associates, firms may eventually have to reconsider how that pyramid works.
There are tentative signs that junior hiring strategies are changing. A Citi survey cited by the Financial Times found that 47% of large law firms had smaller summer associate intakes this year than last, while 49% expect next summer’s class to be smaller again.
It would be premature to attribute that entirely to AI. Hiring decisions are influenced by deal flow, utilisation, economic confidence, individual firm strategy and numerous other factors.
The broader market also remains strong. Large US law firms reported average revenue growth of 12.4% in the first half of 2026, while demand measured by billable hours increased 4.8%. Lawyer headcount grew too, albeit more slowly than in the previous year.
So this isn’t a story about BigLaw suddenly stopping hiring.
The more interesting question is what firms will increasingly be hiring lawyers to do.
What Does This Mean for Legal Hiring?
If technology allows teams to complete certain routine tasks with fewer hours, firms may become more selective about where they add headcount.
That doesn’t necessarily mean fewer opportunities across the board. Demand for high-quality lawyers remains strong, particularly in sophisticated practices where judgement, technical expertise and client management are difficult to replicate. But the profile of the associate firms value could begin to change.
Hiring may increasingly favour lawyers who can contribute beyond execution: associates with strong commercial judgement, direct client exposure, sector knowledge, communication skills and the ability to take greater responsibility earlier in their careers.
That could be particularly significant at the junior end of the market. Historically, large associate classes have supported the leverage model underpinning BigLaw economics. If fewer junior hours are required to deliver the same matter, firms may not need to recruit at exactly the same scale indefinitely — but the lawyers they do hire could be expected to develop more quickly.
For candidates, this changes the career conversation too.
When evaluating a firm, the quality of training, exposure and responsibility may become even more important. A platform that provides a high volume of work is not necessarily the same as one that develops the skills lawyers will need in an AI-enabled market.
And for firms competing for talent, there is another consideration. Associates themselves are becoming increasingly conscious of how prospective employers are approaching technology. A firm’s AI capabilities, training and willingness to give lawyers substantive responsibility could eventually sit alongside compensation, culture and quality of work when candidates compare opportunities.
AI therefore has the potential to influence both sides of the recruitment market: the lawyers firms want to hire, and the firms those lawyers want to join.
The Value of an Associate Is Changing
If technology can perform more of the first-pass research, document review and repetitive analysis that once occupied junior lawyers, the value of an associate increasingly moves towards the things technology cannot easily replicate.
Judgement. Commercial understanding. Client relationships. Communication. Negotiation. Knowing when an AI-generated answer is wrong. Understanding why something matters rather than simply identifying it.
That could ultimately be positive for junior lawyers. For years, one criticism of the traditional BigLaw model has been that highly qualified graduates spend enormous amounts of time performing repetitive work before being given meaningful responsibility. If AI accelerates that process, associates could potentially be exposed to more substantive work earlier in their careers.
But there is a challenge attached to that opportunity.
Many of the repetitive tasks AI is beginning to accelerate have historically been how junior lawyers learned. Reviewing hundreds of documents may not be glamorous, but it builds familiarity with how agreements are structured. Researching a point from first principles develops legal judgement. Working through due diligence teaches lawyers what to look for when something doesn’t seem right.
If technology removes some of that repetition, firms will need to become much more deliberate about how those skills are developed.
The associates who thrive may therefore not simply be those who can use AI fastest. They will be those who combine technological fluency with the technical foundations, commercial awareness and judgement required to interrogate its output — and add value beyond it.
Could Fewer Hours Actually Make Firms More Profitable?
There is another possibility that shouldn’t be overlooked: AI does not necessarily have to make BigLaw less profitable.
Morgan Stanley’s legal leadership has suggested that firms could potentially maintain profitability by using AI to handle more matters while lowering the cost of delivering each one. Instead of maximising the number of hours billed on an individual matter, firms could increase the volume of work their lawyers are able to handle. (ft.com)
That creates a very different economic incentive.
Under an hourly model, completing a task in ten hours rather than twenty can reduce revenue. Under a fixed-fee model, completing the same work more efficiently can increase the firm’s margin.
AI suddenly becomes something firms are financially incentivised to use as effectively as possible.
That could accelerate adoption considerably. It could also affect recruitment. Associates who understand how to use technology intelligently, manage matters efficiently and spend more of their time on higher-value work become increasingly valuable in a model where productivity matters more than simply accumulating hours.
Clients Will Still Pay for Judgement
None of this means clients suddenly want cheap lawyers.
Morgan Stanley has been explicit that it remains willing to pay significant sums for the judgement and talent of the best lawyers.
This distinction matters; AI may put pressure on the price clients are willing to pay for routine work. It doesn’t necessarily reduce what they are prepared to pay for the partner who understands how a regulator will react, the associate who spots the issue nobody else has noticed or the team capable of navigating a transaction when circumstances change overnight.
If anything, the gap between routine legal work and genuinely high-value legal judgement may become more visible.
For lawyers thinking about their careers, that makes the direction of travel increasingly clear. Technical ability remains fundamental, but the associates who stand out are likely to be those who can combine it with commercial judgement, strong communication, client skills and an ability to use technology without becoming dependent on it.
A Different BigLaw Model?
The billable hour has survived predictions of its demise before, and it would be premature to declare that AI will finally kill it.
But something important has changed.
The pressure is no longer coming solely from legal technology companies promising to transform the profession. It is coming from some of BigLaw’s biggest clients, asking why the efficiencies they were promised aren’t yet appearing in their legal bills.
At the same time, firms are paying associates more, investing heavily in technology and competing aggressively for elite talent. US law firm revenues remain exceptionally strong.
Those forces cannot remain disconnected indefinitely.
The most likely future may not be the end of the billable hour, but a market where clients become less willing to pay by the hour for work that technology has made predictable and repeatable, while continuing to pay a premium for complex advice and exceptional judgement.
For legal recruitment, that distinction matters.
Firms will still need talented lawyers. But as the work changes, what constitutes talent may change with it. Commercial judgement, client skills, technological fluency and the ability to take responsibility could become increasingly important differentiators when firms decide who to hire and promote.
For associates, the question therefore isn’t simply whether AI will affect their jobs: it’s whether they are developing the skills that firms, and their clients, will value most as the economics of BigLaw evolve.
