AI, Discovery, and What it Means for the Future of the Billable Hour

AI appears to be challenging the billable hour.

It is.

But not in the way many people think.

The common prediction is simple: AI will make legal work faster, lawyers will bill fewer hours, and the billable hour will collapse. Maybe. But that is not the more interesting story.

The more interesting story is this: the billable hour became dominant because litigation became unpredictable. And much of that unpredictability came from discovery — the long, expensive, adversarial search for facts.

One way to understand the rise of the billable hour is as a shift in economic risk. When the amount of work required to resolve a matter became harder to predict, hourly billing moved that uncertainty from the seller of legal services to the buyer. The lawyer no longer had to price the entire journey in advance. The client paid for the journey as it unfolded.

That made sense in a world where discovery could expand, disputes could shift, and the factual record could take months or years to understand. But it also created a misalignment. The client wanted resolution. The system rewarded effort.

The birth of the billable hour

Several years ago, I took a deep dive into the history of the billable hour. I studied Reginald Heber Smith, whose work at the Boston Legal Aid Society beginning in 1913 helped introduce timekeeping as a tool for law-office management. He then brought it to the newly formed Hale & Dorr, where it became a foundational tool. I also read his book “Justice and the Poor: A Study Of The Present Denial Of Justice To The Poor And Of The Agencies Making More Equal Their Position Before The Law” as well as Roscoe Pound’s 1906 speech to the American Bar Association (ABA) entitled “The Causes of Popular Dissatisfaction with the Administration of Justice.” That speech led to the Rules Enabling Act of 1934 and the Federal Rules of Civil Procedure (FRCP) in 1938.

Those federal rules helped institutionalize broad discovery in federal civil litigation. That discovery gave parties greater insight into the facts behind claims and defenses, but it also created an adversarial back-and-forth that made effort, cost, and duration harder to control. Over time, discovery became one of the forces that increased the cost, duration, and complexity of litigation, even as fewer cases ultimately reached trial.

Almost immediately, in the 1940s, the ABA began promoting the billable hour as a fix to the “economic plight” of lawyers who were falling behind in professional earnings. Before the FRCP, lawyers were billing by retainers and fixed fees. This was embodied in the state bar associations that published catalogs of every conceivable legal service, along with a fee that was enforced among the members—until the Supreme Court ruled that as price fixing in the landmark case of Goldfarb v. Virginia State Bar (1975).

This is where AI becomes relevant. Not because it replaces legal judgment, but because it may help reduce the factual uncertainty that made litigation so hard to price in the first place.

It turns out that lawyers spend a lot of their time, and a lot of case work, focused on the pursuit of evidence, not the application of law. In his eighteenth century “Commentaries on the Laws of England,” Sir William Blackstone observed that “ the experience will abundantly show, that above a hundred of our lawsuits arise from disputed facts, for one where the law is doubted of.”

And our federal discovery rules were penned by Edson Sunderland. He was the law professor of Charles Clark, who was tasked with the overall effort to write the FRCP. Sunderland was a big believer that if everyone knew the facts, they wouldn’t need to go to court. As a result, the rules embodied mechanisms that allowed and encouraged the zealous pursuit of evidence. He was mostly right. He may also have been 100 years too soon.

I say that because only in the past 20 years have we found ourselves emerging from a world of evidentiary scarcity to a world of evidentiary abundance. Mostly due to computers, the Internet, social media, and the digitalization of … everything.

Evidence is everywhere, all at once

Evidence is now abundantly created, and digitally accessible. You might assume that this could make the situation worse, not better. And arguably it did. In the first 20 years or so, the novelty and volume of eDiscovery increased costs. Significantly. I know. I was there. But now the tide is turning.

Today we have tools and methods that facilitate more efficient access to our growing collections of transactional records, written content, and system-generated data (SGD) that capture, and can replay, human-driven activity. And now we see that AI-enabled tools can help lawyers and experts analyze large collections of data faster than traditional linear workflows, provided the process is designed, tested, and supervised.

My prediction is not that AI will diminish the legal acumen of the practitioner, it is that AI will improve access to the underlying evidence to which legal judgment is applied.

If AI can help lawyers and experts find, understand, and test evidence earlier, it may do more than make legal work faster. It may make legal work more predictable.

And predictability is what makes value billing possible.

The billable hour became necessary because discovery made litigation unpredictable. If AI helps reduce factual uncertainty earlier, it may create the conditions for more value-based legal work—and move us closer to the promise of Rule 1: the just, speedy, and inexpensive determination of every action and proceeding.

Closing thoughts: Join the conversation

This is just one piece of the bigger conversation on the future of evidence. As legal professionals, we need to stay on top of emerging technologies.

Let’s continue the discussion on this LinkedIn post.