home  /  library  /  stanford-codex-future-law-firm
Industry research & surveys

The Future Law Firm: Agentic AI and the Restructuring of Legal Services

CodeX — the Stanford Center for Legal Informatics · 2025 · Stanford Law School

The futureBilling & feesManaging partners thinking past the current budget cycle
Why it is on the shelf. The most useful available framing of the structural question, because it is about where legal spend goes rather than about what tools do.

The Institute's reading

The distinction this work turns on is worth carrying around. A copilot is a tool a lawyer uses to do their work faster; the firm keeps the client, the relationship and the fee, and the only question is who captures the efficiency. An autopilot sells the work directly to the end client, and the firm is not in the transaction at all.

The argument is that the shift from the first to the second is underway, beginning with high-volume routine work, and that corporate legal departments will progressively redirect budget from law firms toward technology-first providers for exactly those categories. The mechanism is unremarkable and that is what makes it credible: it does not require AI to become better than lawyers, only good enough at commodity tasks that an in-house team stops sending them out.

For a firm the useful exercise is not to argue with the thesis but to sort the book against it. Which matters are the firm winning on production capacity — volume, throughput, cost per document — and which on judgement, relationship and accountability? The first category is exposed on this analysis. The second is not, and may be worth more as the first commoditises.

Key propositions

  • AI in legal is shifting from copilots that assist lawyers to autopilots that sell work to clients directly.
  • Corporate legal departments will redirect spend toward technology-first providers for routine work.
  • The transition begins with high-volume, well-specified tasks rather than complex matters.
  • The firm’s defensible position rests on judgement, relationship and accountability rather than production.

In practice

  • Sort the firm’s book by whether the value delivered is production capacity or judgement.
  • Work exposed to disintermediation should be priced and staffed differently from work that is not.
  • Client relationships become more valuable as production becomes cheaper, not less.

Where authorities disagree

The most substantial counter-argument is induced demand: that cheaper legal services expand the market rather than shrink the profession, and that most legal need currently goes unmet because it is unaffordable. There is also a regulatory objection — unauthorised practice rules and professional privilege constrain how far a technology provider can go without a lawyer, and those constraints are not obviously about to give way. The Institute's view is that the direction is credible and the pace is very uncertain, and that firms should plan for the direction rather than a date.

AI adoption conciergeorientation · not legal or ethics advice
Happy to dig into it. What would you like to pressure-test from The Future Law Firm: Agentic AI and the Restructuring of Legal Services: one of its propositions, how it applies to your situation, or where it disagrees with the rest of the shelf?