Every six months, somebody publishes another article declaring that AI is about to destroy the billable hour. Every six months, the billable hour declines to die. The articles get a lot of traffic. The pricing model continues unchanged.
This pattern produces a kind of intellectual fatigue. Senior partners read the next “billable hour is dead” article, recognize the genre, and discount the whole question. They have heard this before. They have continued to bill at hourly rates. They will continue to bill at hourly rates next year. The articles, in their view, are written by people who do not understand the actual economics of legal work.
I think the partners are partly right and partly wrong. The articles are written by people who treat the legal industry as one market. It isn’t. Transactional work and litigation are economically different businesses, and AI is going to break them at completely different speeds.
If you want a useful prediction—instead of another generic prophecy—you have to separate them.
The transactional case: collapse is closer than partners think
Transactional work has a property that nobody likes to say out loud. Most of the output is standardized. A merger agreement is a variation on a known template. A capital contribution document follows known forms. A loan agreement has perhaps a dozen genuinely creative provisions and several hundred pages of standard machinery.
The billable hour exists, in transactional practice, because clients have historically had no way to evaluate which hours are “the creative ones” and which are “the machinery.” So they pay for all of them at lawyer rates, and accept that some fraction of what they’re paying is overhead.
AI changes this in a specific way. It does not eliminate the need for the creative judgment. It eliminates the invisibility of the boundary between creative judgment and machinery. When a client can run their own draft through a model and produce something 80 percent of the way to the same output the law firm would have produced, the line between “what I’m paying my lawyer for” and “what I could do myself” becomes uncomfortably visible.
The collapse, when it comes, will look like this: clients with sophisticated in-house teams will start producing first drafts internally. They will hire outside counsel for the final 20 percent—the negotiation, the unusual provisions, the strategic judgment. They will resist paying for the first 80 percent at the same hourly rates as before. Pricing will move toward fixed fees for defined deliverables, with hourly rates surviving only for the genuinely creative tail.
This is already happening with sophisticated corporate clients. It will spread. The pace at which it spreads depends on how quickly mid-market clients build internal AI capability. My estimate: three to five years before transactional billable rates are under serious pressure across the market.
The litigation case: the billable hour is more durable than people think
Litigation is a different animal.
The output of litigation is not a document. It is the resolution of a dispute, achieved through some combination of pleading, negotiation, deposition, motion practice, and trial. The intellectual work is interleaved with the procedural work. The strategic judgment cannot be separated from the production of paper, because the production of paper is itself the strategy.
This matters for billing because the link between time spent and value created is much harder to disaggregate. When I draft a motion to dismiss, I am not just producing a document—I am calibrating tone for a specific judge, anticipating opposing counsel’s likely responses, deciding which arguments to feature and which to hold in reserve. AI can help with parts of this, but it cannot replace the judgment about which fights to pick and how aggressively to prosecute them.
More importantly, litigation involves negotiation with humans whose behavior is also calibrated by hours. Opposing counsel charges by the hour. Their incentives, their pace, their willingness to entertain compromise—all of these are shaped by the same pricing structure. Trying to fix-fee one side of an adversarial human interaction while the other side bills hourly creates immediate strategic problems.
I expect the billable hour to survive in litigation for considerably longer than in transactions. Maybe a decade. Maybe more. The pressure will come from clients trying to cap costs, from third-party litigation funders demanding outcome-based pricing, and from the slow erosion of trust between corporate clients and their litigation counsel. But the underlying economics—pricing the open-ended adversarial process—will resist the change longer than transactional pricing will.
What this means for partners deciding right now
If you are a transactional partner reading this, my advice is uncomfortable. The window in which “we bill by the hour because that’s how everyone does it” is a defensible answer to client pressure is closing. You have, in my view, two to three years to get ahead of this. After that, the firms that have moved to outcome-based pricing in transactional work will be eating your lunch.
If you are a litigation partner, you have more time, but you should not relax. The pressure is coming. It will arrive first as client demands for budget caps and matter-level fee predictability. It will eventually arrive as full alternative fee structures. The firms that figure out how to price uncertainty will outcompete the firms that try to keep billing as if the last decade still applies.
If you are a managing partner trying to make a single firm-wide decision about pricing, I think that’s already a mistake. The two practice areas need different answers. Treating them as one market is the same error the “billable hour is dead” articles make—just from the inside instead of the outside.
A note on what clients are actually thinking
I want to close with the part of this that nobody quite admits at the partner level.
Clients are not going to stop paying for legal work. They are going to stop paying for legal work at *current