Goldman Sachs and Citi Are Telling Law Firms to Bill Less Because AI Is Faster

The billable hour has survived a lot of predictions about its death, but the current challenge is coming from the people who pay it. As reported by the Financial Times and summarised by Seeking Alpha, Goldman Sachs, Morgan Stanley and Citi have told law firms their bills should fall because technology has made routine work faster. For anyone working in or entering law, this matters more than any individual firm's AI announcement, because it changes what clients are willing to pay for. At Metaintro, we track pricing shifts like this because how a profession bills determines how it hires, promotes and trains.

What are the banks actually asking for?

A share of the savings, and evidence that they exist. Goldman Sachs asked law firms how much AI was saving them and expected to share in those benefits. That is a straightforward commercial ask, and it is difficult to argue against in principle once a firm has publicised its own efficiency gains.

Citigroup has gone further into process. The bank started asking law firms to bid for work and explain how much they were saving using AI. Its global head of legal said that if lawyers take less time on a matter because of AI, the expectation is for costs to come down significantly per transaction, and that a different working model based on that approach would probably be in place within a year.

Read that timeline carefully, because it is unusually short for a change of this kind. A year is not a consultation, it is an implementation. And the framing has moved from hourly rates to cost per transaction, which is a different unit entirely. Once a client is buying an outcome rather than time, the internal economics of a law firm change at every level.

Morgan Stanley's general counsel put the consequence plainly, describing law firms' compensation model as extraordinarily unstable. That is a client saying the supplier's business model no longer works, which is a rare thing to say out loud.

Is the legal market actually shrinking?

No, and this is the part that makes the story more interesting than a simple decline narrative. According to Bloomberg Law, law firms recorded 11.7 percent revenue growth in the first half of the year, powered by a rarely seen level of increasing demand, on data from Citi.

Billable hours grew too. Firms across the industry saw 4.2 percent growth in billable hours compared with the same period last year, against a long term demand growth average of between 1.5 and 2 percent. That is roughly double the historical norm. Surging transactional practices handling both megadeals and middle market mergers and acquisitions drove the increase.

So the profession is not being automated out of work. Lawyers are working more hours, not fewer. What is under pressure is the price attached to those hours, and specifically the price of the routine portion. That distinction matters enormously for a career decision, because a shrinking profession and a repricing profession require completely different responses.

Where is the money actually going?

Into costs, which is the squeeze from the other side. Total expenses rose 9.7 percent from the year ago period. Expenses are rising even as lawyers work more hours, driven partly by increased headcount, an ongoing talent battle and the race to adopt artificial intelligence itself.

There is a detail in that data worth pausing on for anyone thinking about progression. Income partners, described as a particularly expensive cohort, are the fastest growing group of lawyers at law firms. That tells you where firms are adding people, and it is not at the bottom.

Put the three numbers together and the picture is coherent. Revenue up 11.7 percent, hours up 4.2 percent, expenses up 9.7 percent, and the largest clients simultaneously asking for the price to fall. A firm absorbing rising costs and rising demand while its biggest customers push for per transaction pricing is a firm that will look hard at which work it does with people and which it does not.

What does this mean for junior lawyers specifically?

This is the pressure point, because the routine work is the training ground. The banks' argument is precisely that technology has made routine work faster, and routine work is what junior lawyers have historically done in volume while learning the profession.

If the client will no longer pay full rate for document review, first draft production or standard diligence, a firm has three options. It can absorb the cost, which the expense figures suggest is unattractive. It can do the work with fewer juniors and more tooling. Or it can change what juniors do. Only the third of those preserves the training pipeline, and it requires deliberate effort rather than drift.

The early signs are already visible in summer associate cuts and in what law school graduates now have to plan around. The same dynamic runs through support roles, where AI legal research is changing paralegal and lawyer jobs at the same time.

The counterweight is that hours are growing strongly, and much of that growth is in transactional work, which is judgement heavy and client facing. Juniors who get onto deal teams are in a better position than those doing volume review, and that has always been true. It is simply now the difference between a career and a squeeze rather than a matter of preference.

Does this change the billable hour permanently?

It is the most serious challenge it has faced, for one specific reason. Previous attempts to move away from hourly billing came from law firms themselves, or from clients seeking predictability. This one comes with a measurement attached. Citigroup is not just asking for a lower price, it is asking firms to explain how much they are saving using AI, which turns the negotiation into an evidence based one.

That is difficult to resist for a profession that has spent two years publicising its AI adoption. A firm cannot simultaneously tell the market it has transformed its efficiency and tell its largest client that nothing has changed about cost.

The likely outcome is not the abolition of the hourly rate but its retreat into the work that genuinely cannot be estimated, with fixed and per transaction pricing taking the predictable remainder. That is already the direction of the stated expectation for costs to come down significantly per transaction. For lawyers, the practical consequence is that hours worked stops being the measure of contribution, and the questions become which work you can price, which you can win, and which only you can do. The wider version of it is already in court, in what an AI law firm winning its first case means for legal careers.

Why are the banks in a position to force this?

Because of concentration and because of what they can see. A small number of very large financial institutions account for a disproportionate share of the transactional work that has been driving demand growth, and that work is exactly what surged, with megadeals and middle market mergers and acquisitions cited as the driver. A client that represents a meaningful share of a practice group's revenue can ask questions a smaller client cannot.

They also have unusually good visibility. Citi's own law firm advisory group is the source of the industry data showing 11.7 percent revenue growth and 4.2 percent hours growth. When a client can see the profession's aggregate economics and its own invoices at the same time, the negotiation is asymmetric in a way it has not been before.

There is a further reason the timing works. The banks are running the same efficiency exercise internally. That is playing out in what Goldman Sachs leadership says about AI and banking jobs, in the bank's argument that AI job fears are overblown, and in JPMorgan monitoring junior banker hours. An institution measuring its own efficiency gains is well placed to ask a supplier to do the same.

What happens to the training pipeline if this lands?

This is the question the profession has not answered, and it is worth naming clearly because it determines what a legal career looks like in ten years.

The traditional apprenticeship in law is economic as well as educational. Juniors learn by doing volume work, and clients pay for that work, which is what funds the training. Remove the client's willingness to pay full rate for the routine portion and the funding mechanism for the apprenticeship goes with it. The work still needs doing, but it stops being billable at a rate that supports a large junior cohort.

The Citi data suggests firms are already responding by adding at a different level, since income partners are the fastest growing group of lawyers. Growth concentrated in an expensive senior cohort while entry level economics deteriorate is a recognisable pattern, and it is the same one appearing across professional services. It produces a profession that is excellent at deploying experience and progressively worse at creating it.

For anyone entering now, the practical implication is that you cannot rely on the institution to build your expertise as a byproduct of billing. The learning has to be sought deliberately, through the matters you volunteer for and the people you attach yourself to, which is precisely why mentoring matters more in the AI era than it did when volume taught the lesson for you. The DOJ losing a quarter of its lawyers shows how quickly a pipeline can thin, and how to start a career as an attorney covers the entry decisions that matter most.

What do the underlying legal career numbers look like?

Strong, which is the part the billing debate obscures. The Bureau of Labor Statistics puts the median annual wage for lawyers at 159,670 dollars in May 2025, with employment projected to grow 5 percent from 2025 to 2035, faster than the average for all occupations. Paralegals and legal assistants had a median of 62,890 dollars in the same month.

Those are not the numbers of a profession in retreat. They are the numbers of a profession whose internal economics are being rearranged. The pressure described here lands on how the work is priced and on who does the first draft, not on whether the work exists. That distinction matters most to anyone deciding whether to enter, because the honest answer is that the career is still there and the apprenticeship inside it is the part under threat. Citi and its peers are not trying to shrink the legal profession. They are trying to stop paying by the hour for work an hour no longer measures. JPMorgan Chase, the third institution named in the reporting, is running the same exercise on its own cost base at the same time.

What does this mean for your career?

Six points, depending on where you sit.

If you are a junior lawyer, get onto the work clients cannot commoditise. Transactional practices are where the demand growth is, and judgement heavy matters are the hardest to price per transaction. Volume review is the exposed category, a point reinforced by the risks AI fake citations have created in court, which shows that unchecked automated output carries its own professional liability.

If you are choosing a legal career now, go in with clear eyes rather than being deterred. Demand is running at roughly double its long term average and revenue grew 11.7 percent in the half. This is not a collapsing profession. It is a profession whose entry level economics are being rewritten, which changes how you should read the route to becoming a lawyer and how to choose your path as an attorney.

If you are a paralegal or in legal support, the repricing reaches you first, since your work is closest to the routine category the banks named. The paralegal salary guide, paralegal interview questions and paralegal resume examples are worth reading alongside a realistic view of which tasks are being automated. Where firms are hiring is shifting too, as big law remote hiring shows.

If you are in house, you now have a template. Three of the largest banks have asked their firms to quantify AI savings and share them. That approach is not restricted to institutions of that size, and the request itself is the innovation. It is worth reading the bank's public argument that AI job fears are overblown against what its own legal team is now asking suppliers to prove.

If you are outside law entirely, watch the mechanism rather than the industry. A client asking a supplier to prove and share efficiency gains is a pattern that will spread to every professional service priced by time, including consulting, accounting, agency work and contracting. The billable hour is simply the most visible version, and the same logic reaching accounting and consulting is what makes human judgment more valuable than throughput.

Finally, understand that the compensation model being called extraordinarily unstable is the same model that funds partner tracks, salaries and bonuses. Instability at the top of that structure eventually reaches every level of it. How the profession is adapting shows up in the hiring surge at law firms and in career lessons from top women lawyers in the AI era.

Related Articles

• Big Law Faces a Hiring Slowdown Amid a Lateral Market Surge
• Why Companies Are Turning to Lawyers for Remote Work Policies
• Attorney Versus Doctor, a Career Comparison
• What It Takes to Get a California Lawyer Licence
• Anthropic's Legal Plugin Triggered a Market Rout and What It Means
• European Banks May Cut a Fifth of Jobs to AI by the End of the Decade
• Inside Jane Fraser's Remake of Citigroup and What It Means for Bank Jobs
• What BCG's AI Revenue Means for Your Next Consulting Job
• Interns Get a Shot at Bank of America While Wall Street Trims Entry Level Roles
• How Nurses and Physicians Move Into Health Law
• Banking Job Titles That Fast Track a Finance Career
• APAC Finance Hiring Surges as Citadel Securities Doubles Its Asia Headcount

People Also Asked

Q: Are law firms losing work because of AI?

A: Not currently. Law firms recorded 11.7 percent revenue growth in the first half of the year and 4.2 percent growth in billable hours compared with the same period last year, against a long term demand growth average of between 1.5 and 2 percent. The pressure is on pricing rather than volume, with major banking clients arguing that bills should fall because technology has made routine work faster.

Q: What are banks actually demanding from their law firms?

A: Evidence and a share of the savings. Goldman Sachs asked law firms how much AI was saving them and expected to share in those benefits. Citigroup began asking firms to bid for work and explain how much they were saving using AI, with its global head of legal saying costs should come down significantly per transaction and a different working model would probably be in place within a year.

Q: Is the billable hour going away?

A: It is under its most credible challenge, but the data does not show it disappearing. Morgan Stanley's general counsel described law firms' compensation model as extraordinarily unstable, while firms are simultaneously seeing hours grow and expenses rise 9.7 percent year on year. The more likely outcome is hourly billing narrowing to unpredictable work while fixed and per transaction pricing takes the rest.

Wondering what you should earn as the pricing model shifts under your profession? Metaintro tracks pay, demand and hiring across legal and professional services so you can see which practice areas are actually growing. Create your free Metaintro profile and benchmark your role against the market.