Jake Heller Interview With Artificial Lawyer: ‘We Never Planned For This’

Last year, Jake Heller, the co-founder of legal research pioneer Casetext, sold the business to Thomson Reuters for $650m in an all-cash deal. At the time he still owned a notable share of the company. Artificial Lawyer caught up with him to hear how life is now and about his views on how genAI will change our lives as well.

First, let’s get the money bit out of the way. When asked about the incredible sale, Heller is modest: ‘We started the company over a decade ago and we were not expecting this. We have never planned our lives around this. I was very happy with what I had before, and the truth is that if you’re not a happy person then getting richer will still not make you happy.’

That said, he’s well aware of what he can do now.

‘Money has made things more convenient. You have people you can ask to do tasks for you,’ he notes, and we’ll come back to the assistant point in a moment.

Did the sum surprise him, which was around 20X revenue? He replies that the way the company has been growing since the deal, and how it was growing just before the purchase, leads him to believe this was not so surprising. Also, you have the ChatGPT effect, which was perhaps at its strongest point over the summer of 2023.

‘Lawyers could tell that this would change everything. Managing partners wanted to know what their firm’s AI strategy was. And we were able to ride that wave,’ he adds. ‘So, overall we made our own luck, but yes, it was lucky [in relation to the timing of the deal].’

We then have a chat about how the deal came together, which we cannot share the details of here. However, it can be said that if Heller’s hectic diary last year had worked out in a different way he may not have met with Thomson Reuters at the right time and rather than being bought by them, Casetext would have completed another major funding round and still be independent today. Plus, inevitably Thomson Reuters, as has happened with its rival LexisNexis, would probably have built a genAI capability all by itself.

But, they did meet, things did move forwards – very quickly – and as they say, the rest is history.

Heller then adds that timing really is essential to everyone’s company – especially a tech company. In their case, they were early with AI, using ‘traditional’ ML / NLP and ‘we assumed that the AI would get better’. And it did get better, but there were not the massive leaps some had hoped for.

‘For a long time the tech was not where we needed. But we stuck it out,’ he says. Then LLMs got really good, really fast. ‘The take-away is that sometimes it takes a decade for the tech to reach the right level and then suddenly everything becomes possible,’ he adds.

Not that Casetext only discovered LLMs in late 2022, they’d been exploring this area of AI tech for some time, in fact since 2018, primarily for search as it could help find things even if you didn’t use the exact words that matched to an aspect of a case. But, even then with this approach, before 2022, the capabilities were just not there yet. So, this really is a story of timing and sticking with it until the tech matured to where it could really change things.

And of course, in a parallel world this maturation may not have happened, or perhaps the company languished because of the lack of AI improvement. After all, we have seen this happen in other sectors (and also in legal tech itself) where tech promises didn’t turn into reality and customers grew tired of waiting for the ‘great leap’.

But, as noted, for Casetext things did work out. They’d developed expertise in using genAI for case law research. The buzz in the legal and tech worlds was at a deafening crescendo. Companies such as Thomson Reuters – whose foundations are firmly in legal research – were determined not to be left behind with this new wave of AI. There was a feeling in the air that now was the time to do something….perhaps now was a time when you had to do something. And so it happened.

Talking to many other folks in the legal tech world, now that the dust has settled, most say that this was almost the kind of deal that proves the exception to the norm. It was a special moment in legal tech history and such a deal between two companies may not come again for a long while. It just happened to be Casetext and Thomson Reuters – the stars had aligned.

The Future

So, what now? CoCounsel, which Casetext created as an AI assistant and pre-dates the deal with Thomson Reuters, goes from strength to strength, and Thomson Reuters is now going to use the approach of a genAI assistant in all its other sectors.

And Heller is quick to point out that CoCounsel could already do more than legal research before the sale. ‘It could review documents and contracts, summarize documents, draft correspondence, help you prepare for a deposition, and much more. There’s sort of a characterization that we were just about legal research, but that’s not quite right,’ he noted.

‘It’s a really exciting vision,’ says Heller of where things are heading with CoCounsel, then adds that this is the beginning and that genAI assistants will be doing ‘all the research, the first draft of documents, summaries, and will read all your emails’.

He underlined that we are at ‘a tipping point’ and that as with iPhones all lawyers will be using this technology simply because it’s so clearly useful.

‘It will be impossible to avoid, like email. It will be ubiquitous,’ Heller explains.

And this follows through to the rest of the world, not just lawyers. In the future, assistants will help with everything they can.

Heller puts it this way: ‘You know when you see some people, and they are so neat despite their schedules, and their homes are perfect, and everything about their life is perfect? How do they do this? They have real life assistants,’ he observes and mentions politicians and movie stars as an example.

‘We will be the same, (i.e. regular folks),’ he says. ‘This will be democratised. We cannot employ all these people, but we can have AI.’

He then adds that in terms of the physical dimension eventually there will be robotics to help.

‘Everyone wants to be the person who is assisted, it’s just too expensive right now,’ he concludes.

Now, that truly is looking to the future, and the robotics part is likely a very, very long way off, but the main argument makes sense. Throughout the history of technology, tedious tasks have been improved via machinery, then software, and sometimes both.

An assistant is someone who does things you’d prefer not to do. That might be summarising the EU AI Act for a law firm partner, or it might be a more manual task at home.

Tech’s journey is to slowly remove the need for human work for tasks we don’t want to do, or that we could do but we’d rather focus on something more high value and rewarding.

From Heller’s point of view, the Casetext/Thomson Reuters deal is part of this continuum of evolving technology taking over these assistive roles. In his case, he has been lucky enough to become a direct part of this continuing story.

And if you found this interesting, then come along to Legal Innovators California, June 4 and 5 in San Francisco, where subjects such as genAI assistants and much more will be explored with an incredible group of speakers!

Legal Innovators California conference, Jun 4 + 5, San Francisco

Day One is focused on law firms, and Day Two on inhouse and legal ops. 

We have many great speakers attending the event, along with a group of pioneering legal tech companies and service providers – you can see some more about our speakers here. It will be two great days of education and inspiration! Join us and get ahead of the curve on all things legal tech, innovation and legal AI! 

For ticket information, please see here. Don’t miss out on what will be a great event.