Wall Street Banks Demand Big Law’s AI Savings — Is the Billable Hour Broken?

Corporate legal teams now use AI themselves—so they know how fast document review, research, and first drafts can move. That knowledge is colliding with record Big Law rates. In early September 2026, reporting on Wall Street’s largest banks described a blunt ask: if AI is compressing legal work, clients want a share of the savings—not the same invoice for fewer hours.

Wall Street’s AI savings audit

According to Modern Counsel and The New York Post (citing Financial Times reporting), Goldman Sachs, Citigroup, and Morgan Stanley are among the institutions pressing outside counsel on AI-driven efficiency.

Citigroup’s global head of legal, Adam Meshel, told the FT that if hours on a matter drop because of AI, the bank’s expectation is that costs come down significantly per transaction. Morgan Stanley general counsel Eric Grossman went further: the associate-hour foundation of mega-firm compensation looks “extraordinarily unstable” as AI alters the revenue base—even while the bank remains willing to pay premium rates for judgment and expertise.

The information asymmetry that once protected opaque hourly invoices is shrinking. Deloitte survey data cited by Modern Counsel shows most legal departments are already deploying or embedding AI. When in-house teams can draft and review with the same class of tools, large associate-hour line items for routine work become harder to defend.

Why “time” no longer equals “value”

Industry economics analysis from Kallam AI frames the problem as arithmetic, not culture:

• Legal AI tools can compress document review timelines by roughly 60% to 70%; NDA drafting can be up to 70% faster with assistance.
• In one high-volume litigation example, a complaint-response system cut associate review time from 16 hours to about 3–4 minutes.

Under the classic model, revenue = hours × rate. Compress the hours and—unless rates or scope are redesigned—the invoice shrinks even when the client receives a faster, often better-timed result. A firm that invests in tools that turn ten hours of work into ten minutes can be penalised for efficiency if it still bills only the surviving minutes.

That is the pricing paradox of the AI era: productivity gains under pure hourly billing transfer value to the client by default, unless the firm moves to pricing that rewards output, outcome, or capacity.

From billable hours to output and outcome pricing

Firms and buyers are already testing new architectures. Thomson Reuters describes growing interest in alternative fee arrangements (AFAs)—flat fees, capped fees, blended rates, and hybrids—as clients demand predictability and value alignment, while AI makes scoped work more forecastable.

On the firm side, the UK’s Law Society Gazette reported that north-west firm IMD Solicitors is moving away from hourly charging toward an output- and outcome-oriented model, with tiered service options (scope, pace, access, senior involvement). Managing partner Marcin Durlak’s framing matches the Wall Street debate: if technology delivers an excellent outcome in less time, efficiency should not automatically mean a lower fee—clients buy expertise and results, not units of time.

Fixed and outcome-linked fees change the incentive map:

• Fixed fee for a defined scope (for example, an acquisition workstream): the firm can use AI to deliver faster, protect margin, and give the client budget certainty.
• Capped hourly: savings may still flow to the client as hours fall below the cap—capacity expansion becomes the firm’s main AI dividend.
• Outcome / value-based: fees track results rather than inputs—harder in uncertain litigation, more natural in transactional and productised work.

Wall Street’s push for competitive bidding and quantified AI savings accelerates the same shift: firms that can show lower delivery cost and protect judgment-priced work will win RFPs; firms that only speed up the same hourly invoice will not.

What law firms should do now

1. Measure comparable matters, not every AI minute. Banks do not need a perfect second-by-second AI ledger. They can compare similar deals on total hours, staffing mix, turnaround, and cost—and put those metrics into competitive bids.
2. Separate compressible work from judgment. Research, first drafts, and large-scale review are where clients will demand AI-aware pricing. Strategy, negotiation, and accountability still support premium fees.
3. Price the model you actually run. Under hourly billing, AI is a revenue-compression risk unless capacity is redeployed. Under fixed fees, the same tools are a margin lever. Pick AFAs deliberately by practice area.
4. Be ready for transparency. Outside-counsel guidelines and RFPs increasingly ask what AI changes about delivery cost. Prepare a clear story: tools used, human review loop, and how fees share efficiency gains.

Implications for Hong Kong and Asia counsel

Sophisticated buyers in Hong Kong and across Asia are watching the same playbook. Global banks and corporates that audit US and UK panels will export expectations on AI savings, AFAs, and matter benchmarking. Local firms that adopt governed, matter-scoped AI for legal research and contract review—without abandoning data control—will be better placed to answer those RFPs than firms that treat AI as a silent associate-hour multiplier.

Bottom line

Wall Street’s demand for a share of Big Law’s AI savings is not a passing procurement skirmish. It exposes a structural crack: when delivery time collapses and rates stay at record highs, time stops being a credible proxy for value. Firms that survive the next phase will price outputs and outcomes where AI compresses labour—and reserve hourly or premium rates for work that still requires human judgment.

Next steps: Explore legal research and contract review, or sign in to test matter-scoped workflows.

Sources

• Wall Street wants its share of Big Law’s AI savings — Modern Counsel
• Wall Street banks tell Big Law to cut fees as AI speeds up legal work — New York Post
• AI for Law Firms: What the Economics Actually Look Like (2026) — Kallam AI
• From billable hours to value-driven legal services — Thomson Reuters
• Firm vows to scrap billable fee model and charge based on output — Law Society Gazette