Skip to content
Legal Practice Management

Pricing AI-driven legal services: It’s not a question of cost recovery

William Josten  Senior Manager, Enterprise Content - Legal, Thomson Reuters Institute

· 8 minute read

William Josten  Senior Manager, Enterprise Content - Legal, Thomson Reuters Institute

· 8 minute read

When determining how to price legal services that have been augmented by AI, law firms need to be careful to not get caught in the old ways of thinking

The legal market is rife with questions about how legal services will be priced as artificial intelligence (AI) becomes a more dominant force in how legal work is performed. Given the data that the Thomson Reuters Institute analyzes and our expertise in the economics of law, hardly a week goes by without someone posing the question: “What impact is generative AI (GenAI) having on law firm billing?”

The truthful answer is not much…yet. The biggest reason for this is because so much of this technology is new. Recall back to early-2023 and how infrequent conversations were around GenAI — then, an idea that was largely synonymous with ChatGPT, prior to the proliferation of large language models and tools we’ve seen since. It wasn’t until just a little over a year ago that the topic really burst onto the stage once ChatGPT was launched in November 2022. Product offerings specific to the legal market followed closely behind but were really only fully available in the later part of last year.

Even assuming that those products were immediately adopted by some lawyers as soon as they became available and those lawyers used them exclusively starting in late-2023, the impact still has yet to register. And given the way legal billing works, any November 2023 work would have been billed in December or maybe even January of this year. Factor in client payment cycles and the very beginnings of the impact of that work on law firm billing would only begin to be seen in law firm collections by the middle of Q1 2024 at best. And even assuming such an edge case, the impacts would not be widespread enough to be able to gauge any market-wide trends or developments at this point.

Of course, that doesn’t mean that we’re not looking for them or that we don’t have suspicions about what will likely happen. I’ve previously written about the potential impact GenAI might have on how law firms measure productivity, and the topic of whether GenAI will lead to greater use of alternative fee arrangements — Spoiler alert: It will — merits scrutiny in its own right. Which brings me to this new body of work the Thomson Reuters Institute is developing around monetizing GenAI and pricing AI-driven legal services.

To be sure, however, law firms must recognize that GenAI remains largely in a definitional period. Law firms and clients are still exploring exactly what GenAI is, what use cases it will most appropriately help, of what risks users need to be cautious, and more. No one is certain of the answers to any of these questions yet, but most legal industry observers are in agreement that the work being done today will lay the groundwork for what an AI-driven future looks like for the profession.

Recovering the costs of AI technology

To that end, I wanted to address another question I’ve been fielding frequently, specifically: “How are law firms planning to recover the costs for their new AI technologies?”

I would like to say, first, that I disagree with the framing of the question.

Lawyers are users of precise language — the words we choose are chosen for a reason, whether to convey a very particular meaning or because they are terms of art. The phrase cost recovery is a good example. For more than 30 years, law firms have engaged in the practice of recovering costs they ostensibly incur on behalf of their clients — costs that range from online legal research services to photocopies to long distance phone calls.

The costs of advanced AI technology should not be lumped into this category.

First, grouping AI in with items like long distance phone calls and photocopies applies a 30-year-old mindset to a brand-new technology. The American Bar Association (ABA) issued the controlling ethics opinion on the question of cost recovery in 1993. ABA Formal Opinion 93-379, title Committee on Ethics and Professional Responsibility Opinion – Billing for Professional Fees, Disbursements, and Other Expenses states as follows (emphasis added):

Perhaps the most difficult issue is the handling of charges to clients for the provision of in-house services. In this connection, the Committee has in view charges for photocopying, computer research, on-site meals, deliveries, and other similar items. Like professional fees, it seems clear that lawyers may pass on reasonable charges for these services… the lawyer is obliged to charge the client no more than the direct cost associated with the service (i.e., the actual cost of making a copy on the photocopy machine) plus a reasonable allocation of overhead expenses directly associated with the provision of the service…

It is not appropriate for the Committee, in addressing ethical standards, to opine on the various accounting issues as to how one calculates direct cost and what may or may not be included in allocated overhead… Any reasonable calculation of direct costs as well as any reasonable allocation of related overhead should pass ethical muster… It is impermissible for a lawyer to create an additional source of profit for the law firm beyond that which is contained in the provision of professional services themselves. The lawyer’s stock in trade is the sale of legal services, not photocopy paper, tuna fish sandwiches, computer time, or messenger services.

Let’s ponder the implications, for a moment, of treating advanced AI technology like the items contemplated in the ABA’s opinion. From the start, the opinion could be read as limiting the ability of the law firm to use AI technology as a source of profit in its own right. Indeed, profit-making would be limited only to the “provision of professional services themselves.” And certainly no one would argue that AI is as valuable to the client as photocopy paper or a tuna fish sandwich.

It’s anyone’s guess as to whether the ABA would treat AI with a similar result should the question of recovering costs for AI rise to its attention. The question need not ever be posed, however.

What do outside counsel guidelines say?

For the most part, clients have begun to largely obviate the question of cost recovery on their own through the implementation of outside counsel guidelines (OCGs). Law firms that choose to add their shiny new AI tech tools to their list of line-item disbursements risk having such tools treated similarly by OCGs. After all, the client would only be following the law firms’ lead in defining the value of the tools.

However, none of these scenarios needs to be given an opportunity to become reality. It is up to law firms to determine how they choose to define the value of their new technologies. If law firms choose to focus on the value of the output — the speed of the result, the breadth of access to knowledge, both in word and deed — then clients will likely follow suit. Clients will also likely have a hard time arguing that a more comprehensive and faster answer to their legal issues is equivalent to a photocopy.

That’ not to say that recovery of some costs is totally out of the question. I recently spoke to a gathering of highly experienced pricing professionals who were in full-scale agreement that generally AI should not be treated as a disbursement. However, they also agreed that if a particular technology had to be purchased for a particular matter or on behalf of a particular client, and it did not have broader application within the firm, such a purchase could reasonably be billed back to the client. However, in most cases in which the technology has broad application across the firm, these experienced pricing pros agreed that billing it back to the client would not be a wise approach.

Indeed, one participant likened it to law firms adopting laptops. GenAI is simply the next evolution of necessary technology, and no firm would have considered applying a per-minute or per-login charge to a client’s bill for every time a lawyer used a laptop when working on a client’s matter.

The impact of GenAI on law firms’ billing practices will likely be in an evolutionary state for quite some time to come. It is important, however, to begin that process on the right foot. Or, perhaps more aptly, not on the wrong foot.

Those law firms looking to make meaningful use of AI in the future should be cautious to avoid applying antiquated business mindsets in these early stages. Those who do choose to treat AI as a disbursement may find themselves as successful with that approach as those still trying to bill their client for a tuna sandwich.


You can find more on Pricing AI-driven legal services here.