Four Reasons Why Lawyers Use Legal AI

Why do lawyers use legal AI? There are four main reasons and each has a different goal, which in turn changes the calculation of ROI. So, here they are, considered from law firm and inhouse perspectives where applicable.

Reduce Unbillable / Low Billable Work – this is the primary reason for large commercial law firms that are in high demand to use legal AI tools. These businesses do not want to reduce billable time, but they will use AI to crush those tasks that clients won’t pay for, or only pay a small amount. Law firms in this position often struggle to find much ROI – which is no surprise at all. If the goal is to apply the efficiency gains of automation (i.e. use AI to perform tasks you normally do manually) only to the lowest value fringes of your business then you can’t be surprised when you see a relatively small return on investment. (Inhouse clearly has no connection to this approach.)

Performative / Marketing – it’s fair to say that plenty of law firms that have experimented with AI since 2023 have done so for performative and marketing reasons – with the messaging designed both for the internal and external audience. The message is: ‘Look, we’re using AI.’ In such cases, you’ll see phenomena such as rolling out an AI chatbot in order to help redraft emails, or help with a bit of brainstorming. In such scenarios there is really little to no ROI at all in clear financial terms. However, one should not dismiss marketing benefits. Law firms spend millions each year on sending ‘the right message’ to the clients, performative AI use can become part of this. Does AL celebrate this? Not at all. But, this list is meant to include what is present in the market. (While some inhouse teams have brought in AI systems to ‘experiment’, there is little reason for inhouse lawyers to go down the performative route.)

Expand capacity (Law firms) – some firms – although not many – are using AI to help expand capacity even when it absorbs valuable billable work. Why do that? The reason is that these are law firms that are very busy, but which don’t want to hire additional staff to cope. As the firm is ‘fully engaged’ then even if you’re using AI in a way that loses some billables, you’re getting work done more quickly and thereby keeping the clients happy. As there is a lot of demand – at present, at least – using AI in this way keeps operating costs down, yet the firm is still billing at maximum capacity. Plus, more clients get more work done more quickly. Again, it’s hard to show ROI here as every lawyer, or at least every practice group, will be cannibalising different types of billable work. Also, as the firm is earning about as much as it can on a time-based approach, how can you show improvement?

Expanding capacity for inhouse teams, which we’ll include under the same category, is also a very real reason for using AI. In fact, it’s probably the number one reason for inhouse lawyers. Naturally, there are no billable targets, there is just an endless stream of things to handle, some high value and complex, but a lot is what we can call ‘high frequency / moderate complexity’ – i.e. the regular contract reviews or the day-to-day compliance tasks. AI helps a lot here and adds real efficiency that can be measured very clearly by an inhouse team. For many companies, AI is used perhaps without much workflow design in mind, but because time to value is so rapid, most inhouse lawyers are just glad to have this extra assistance in their work. Of course, we then hit the ‘AI lets you do more complex work’ challenge. Broadly this is true….if your company has enough of that complex work for the team to handle. But, given that the world gets ever more complex, regulated and risk-filled, then there may indeed be headroom for that to happen, at least at some companies. Either way, there is clear ROI to this approach for inhouse teams.

Redesign Workflows for Maximum Impact – and finally there is the approach that truly makes the most out of legal AI, which treats AI with maximum consideration and brings it to bear not as an ‘add-on’, but as a core component.

AI is used here as a central element in reworked and redesigned workflows. The law firm, or inhouse team, takes a strand of work and says: ‘How can we make this as efficient as possible? How do we squeeze every ounce of benefit from using AI here? How do we map things out to streamline processes knowing we have an incredibly powerful AI engine in the middle of it? How do we find and clean and connect key data to this process? What quality checks do we need and where do they go? Who looks after this? How does it connect to the clients, and the rest of the business?’ And more questions. But, as they are unencumbered by old ways of seeking to generate profit via time, or for inhouse have no time targets at all, this is all upside.

If a law firm is using a fixed fee then they will be able to find ROI very quickly and easily. If not, it will be very hard. For inhouse teams, likewise, they will find ROI very easily. The throughput can be measured. The human input for each matter running through the AI-centred workflow can be observed and quantified. Speed of completion can be measured. And more.

In truth, this is the only really significant reason to use AI in the legal world, all the other ones above have limited impact, and attempts by law firms to find some kind of metric to show value is unlikely to succeed, or if there is a number provided then it will be underwhelming, if using the old approach.

AI applied to work focused on language is one of the biggest technological leaps the legal world has ever witnessed. We cannot be surprised that it therefore demands some rethinking of old approaches to maximise its value. A jet engine attached to a horse and cart is not likely to produce great results, for example, (see: The Three Legal AI Models.)

Very few law firms are doing this deep and broad redesign. The legal businesses that are truly embracing this include the New Model, AI-first firms, which are starting with a ‘blank page’ and designing legal workflows with technology at their heart. That said, they know they need legal experts to design and maintain these systems, and to work with the clients as lawyers have always done. But, they’ve escaped the structural trap of the current model. In some cases, such firms are building deeply connected links into inhouse teams, which are sharing their knowledge to help make the process even more efficient and more powerful.

Plaintiff / claimant law firms are also embracing this approach, especially in the US, because they only get paid once a matter is won. Any work, whether legal or bureaucratic, before the win is simply a drain on the firm and a challenge for a limited number of lawyers.

Inhouse teams can also embrace this on their own, whether an external party helps them or not. Rebuilding workflows to drive efficiency is a clear win-win for the business.

And those larger commercial law firms that have elements of their practice that are already often on a fixed fee can also do this. In fact, if you operate a fixed fee stream of work and you’re not leveraging AI to its fullest then you are running that business line less profitably than is possible.

–

Bonus reason – Avoid using lawyers – there is one other reason to use legal AI, although few lawyers will pick this one, and that’s to avoid using lawyers at all. I.e. to crystallise data and processes, with an AI engine at the heart of the ‘machine’, to produce reliable legal outputs to clients without the need to engage a lawyer at all. At present and because of hallucination risks, operating this ‘no-lawyer’ model is very risky. That said, people will use it if they cannot afford a law firm, or simply don’t want to ask a lawyer even if they can afford it. In time, such approaches will get better and the ROI will then be huge, i.e. once the risk of inaccuracy has been crushed. But, we are not there yet.

—

Conclusion

There are many reasons to use AI in the legal world. It all comes down to what you are actually trying to do. Any lawyer using AI needs to ask a simple question: why am I using AI? What is the economic benefit to me and my business? How does this help my clients, whether external or internal? If those answers are not clear, then you may need to go back and think things through until you can come up with an answer.

–

Richard Tromans, Founder, Artificial Lawyer, Jan 2026