As corporate legal departments and law firms everywhere scurry to assess how they can adopt artificial intelligence and improve efficiencies, eight-year-old legal AI firm Luminance is positioning itself at the vanguard of the industry’s latest and most critical technological shift.
Launched in 2015, the company’s rapidly-expanding law firm customer base already includes more than a quarter of the top 100 law firms, including Slaughter and May and White & Case. Mayer Brown, Herbert Smith Freehills and Clyde & Co are among the newer additions. It also works with major corporates such as Koch Industries, Tesco, Lamborghini and Liberty Mutual.
Now its technology is helping with not just document navigation, but even contract negotiation, helping teams gain the advantage around the deal table. On the back of such evolution, the firm in 2022 grew its customer base 10-fold—its U.S. base growing 140%.
Law.com International spoke to Luminance’s CEO, Eleanor Lightbody, about how her company is meeting its moment, why ChatGPT is only ‘dinner party’ smart, how teams are being asked to “do more with less”, and why, despite staggering developments in AI, humans will always be necessary in the business of law.
Answers have been edited for length and clarity.
Luminance has grown rapidly over the past few years. What do you attribute that growth to?
[In-house] teams are being asked to do more with less. So our technology helps drive efficiencies. For example, helps automate negotiation for those low level, high volume contracts. And as we’re being asked to do more with less, we have teams that are saving money on outside counsel, such as Featurebase, they saved 50% on outside counsel.
Our product is tailored to in-house teams, but it has a much broader application, and that’s why we’ve seen such success and growth. We sell to the legal teams, who can drive efficiencies, cut costs and help with the bottom line, and have more time to do the jobs they’re more interested in, the value-add, the business-enablement part. It’s AI wherever computer meets documents.
How exactly can AI assist with contract negotiation?
Customers are using us, for example Hitachi and Yokogawa, uses the AI to do the first pass of a review for any of the low value high volume contracts that come through, such as NDAs, sales agreements, and employment contracts. Our AI conceptually understands what’s normal for the business, and then will adjust, highlight and flag parts of clauses that are contentious or need further review.
This negotiation piece is really key in this day and age as it helps retain institutional knowledge. As a company, you will know how how you negotiated these different types of contract.
The AI gives insights into what you’ve agreed to in the past and what you might not have. Lawyers and in-house teams can find comfort in the fact that no stone is left unturned.
What do you make of the whole GPT ‘movement’?
I’m so excited. My grandmother called me up and was like ‘what’s this ChatGPT, oh I think I suddenly get what your AI does!’. I said, ‘well, we’re not ChatGPT, but yes finally you can see what [we’re doing]!’
It’s made it much more tangible to anyone, the benefits of it. Lawyers, historically, have been quite sceptical of AI, but that has changed massively, and AI has managed to open up those doors.
But how does Luminance differ from ChatGPT?
ChatGPT is a generalist. It’s been trained over the whole internet. At Luminance, we’re specialists. We’ve been using large language models before most people even knew about large language models. We call ours the legal pre-trained transformer. Our AI has been trained over specific legal documents and verified by legal specialists. That’s key. That’s where the differences are.
In the legal environment, if you’re using AI, you need to know it’s accurate, or that it’s explainable. And that if it doesn’t know the answer it tells you it doesn’t know the answer. The way Chat GPT has been built is that it sounds plausible, but it has been built to give you an answer come what may. Lawyers don’t want that.
The way I describe it is [like being at a] dinner party. I might catch up on the news, read Apple News, and sit at a dinner party and sound quite plausible. Whether I’m factually accurate, I don’t really care, because I’m at a dinner party, I’ve had a glass of wine! If I’m sitting next to a lawyer, I’ll be very accurate, and that’s the difference. We’ve trained [our AI] to be very accurate on legal content.
Will robots takeover, and are human roles in law under threat?
As it stands, lawyers don’t have enough time to get through the work they want to get through. I think [AI] will help today with retention and allowing teams to get through work faster. In the future, I think it might change the make-up of a role. But you will always need a human to do the creative thinking around it. To do the advice and strategy around it. To have that human relationship part.
Do I think AI will automate low level tasks? Yes. That’s exactly what my vision is for Luminance. Do I think that there will need to be humans involved at some level, shape or form at some point in the loop, absolutely, because you always need a human to be advising or talking through the outputs.
I wonder if we’ll see the emergence of legal operations teams, or new roles to accompany the AI technology. It’s inevitable that some roles will be redundant. But humans are very good at finding other applications. So I’ll be interested to see what then emerges from that.
Could lawyers just become operators of legal tech systems?
I think we see a lot of lawyers already are. Slaughter and May, White & Case, they use our tech on almost a daily basis. I think lawyers that don’t will be accelerating that. It’s going to be part of all of our daily lives.
Lawyers are sceptical, they’re analytical. They’ve been historically quite slow to change. Do I think [these changes] will happen overnight? No. We’re still probably quite far. But if they’re not [adopting] technology they’re probably going to be left behind.