Our friend Damien Riehl stopped by to talk with Marlene Gebauer about two big happenings at vLex. Riehl unveiled exciting advancements in vLex’s AI-powered legal research platform and shed light on vLex’s commitment to streamlining legal workflows and reducing the need for extensive prompt engineering.
One of the major developments is the enhanced document analysis feature. Users can now upload legal documents, such as complaints, and vLex’s AI will automatically extract key information including claims, facts, parties involved, and potential legal defenses. This eliminates the tedious manual process of reviewing and analyzing documents, saving lawyers significant time and effort. Additionally, the platform suggests relevant legal research questions based on the document’s content, further expediting the research process.
vLex’s advancements directly address the growing concerns surrounding prompt engineering in legal tech. By automating key analytical tasks, the platform empowers lawyers to focus on higher-level strategizing and client interactions, rather than spending hours crafting the perfect prompts for AI tools. Riehl echoes the sentiment of OpenAI’s Sam Altman, believing that successful AI integration should render prompt engineering obsolete. He acknowledges that the option to fine-tune prompts remains, similar to Boolean search techniques, but emphasizes that vLex aims to make it a choice rather than a necessity.
The potential impact on the legal industry is substantial. Clients, especially large corporations, can leverage vLex’s capabilities to analyze past legal actions and assess the value provided by their law firms. This transparency could lead to a shift from billable hours to flat-fee arrangements, incentivizing efficiency and cost-effectiveness.
Further amplifying vLex’s potential, the company welcomes Daniel Hoadley, a renowned legal tech expert, to lead their research and development team. Hoadley’s expertise in data science and large language models promises exciting advancements in harnessing the power of vLex’s vast legal document database. With a robust roadmap of projects, vLex’s is poised to continue pushing the boundaries of legal technology and shaping the future of legal practice.
Listen on mobile platforms: Apple Podcasts | Spotify | YouTube
Contact Us:
Twitter: @gebauerm, or @glambert
Threads: @glambertpod or @gebauerm66
Email: geekinreviewpodcast@gmail.com
Music: Jerry David DeCicca
Transcript
Marlene Gebauer (00:00.398)
Welcome to the Geek in Review. We have a special news announcement that we’re bringing to you. So breaking news, we have Damian Real here from Vlex and he’s got some exciting news to tell us. So Damian, Damian, what’s going down?
Damien Riehl (00:15.734)
Two really exciting things that we’re doing. Number one was regarding the product. We have some really cool features that we’re releasing. And then number two, we have a great hire. So we’re really excited about the stuff that we’re building and the stuff that people were hiring to build other cool things. Let’s do the stuff that we’re building first. So we, in the past, have been able to, you could ask a legal question.
Marlene Gebauer (00:31.246)
What are we going to do first?
Okay.
Damien Riehl (00:39.19)
And we would provide an answer to that legal question in about two minutes. And it will include non -hallucinated cases, statutes, and regulations. We’ve had that now for many months. The second skill is that if you want to compare jurisdictions, say a 50 -state survey or a multiple -country survey, you could be able to do that multi -country survey. Number three, if you wanted to be able to build an argument, that is, make a more argumentative one to find the other side of an argument, you could be able to do that. We’ve had that for a bunch of months.
But what’s really new is we’ve built out in a big way our document skill. So you can upload one or many documents. And then once those documents are uploaded, we could be able to do really fun things with those documents. So for example, for litigation workflows, I could be able to take a particular case like the New York Times and the system, Vincent AI will be able to say, oh, that looks like a lawsuit alleging copyright infringement. It will know it’s a complaint. And then it will programmatically be able to say what it will do for every complaint.
That is extracting all the claims, extracting all the facts, extracting the relief, figuring out what parties are in the lawsuit, finding legal defenses. This is the work of the lawyer and be able to say, what questions should I ask my client? It does this all out of the box, no prompting necessary. Find your documents out of 800 million dockets and documents. You could do that here and then find legal research questions in green that are customized for this particular case. What are exclusive rights for copyright holders and AI and ML?
So really all these things are right out of the box. And we think that this might be the first.
Marlene Gebauer (02:07.726)
Oh, I see. So you don’t even, do you have to click on the button or is it all like down if you just scroll down?
Damien Riehl (02:10.262)
Yeah, so you click on the button, click on the button and then say I want to extract and analyze all the claims. If you do that, and then here’s copyright infringement. Here’s the governing law. It’s federal, not state. Here are the facts for that claim and here are the parties for that claim. We do that analysis for every single one. Timeline of facts. Here’s the analysis for that. All the parties involved. Here’s all the rules for each of the holding companies, et cetera. What is the relief? They’re looking for money. They’re looking for an injunction. They’re looking to destroy Chachi PT.
What defenses? Maybe you want to argue fair use. Maybe you want to be able to argue other things. What questions should I ask the New York Times? What questions should I ask OpenAI? What questions is my legal research question? And now here’s 45 cases, statutes and regulations and administrative decisions and the memorandum that you can see here. Again, I don’t need to prompt at all. This is all right out of the box. So this is a, this is, we think that.
Probably the future where you as a user don’t have to be a prompt engineer, we’ll just tell you all of the cases that are like yours. We’ll give you all of the claims. We’ll give you all the facts. And you as a user are really just able to be able to just do your work like God intended.
Marlene Gebauer (03:21.23)
So a couple of things, I mean, this is really a big deal because I mean, there’s been a lot of discussion about prompting and the challenges of prompting and how that sometimes is a deterrent for people using the systems. And you have basically made this easy for people. You don’t really have to think about crafting something like that. Yeah.
Damien Riehl (03:45.878)
That’s exactly right. When Sam Altman was on stage, I saw him at a talk and someone asked him, what do you think of the ascendance of prompt engineering as a field? And Sam Altman said, ugh, I hope that’s not a thing in a year. He said, because if we’re doing our job correctly, you won’t have to be a prompt engineer. You just ask a question and we’ll give you the answer. So in that way, I agree with Sam. I think that if I’m doing my job correctly, we at Vlex will be doing all the prompt engineering for you. And you as a user just do the legal things that you do so well.
Marlene Gebauer (04:14.542)
But, you know, the librarian in me has to say, it’s like, well, you know, I’ve, I’ve seen these like, you know, if you want, if you want to, you know, this is what you want. You know, it’s like, here’s, here’s the, the, the can search. Um, and sometimes they work and sometimes they don’t work. So, I mean, do you have the flexibility to kind of make adjustments if, if it’s not exactly what you’re looking for?
Damien Riehl (04:36.47)
Yeah, if you want to be a prompt engineer, absolutely. You can go ahead and do that prompt engineering. And if you’re not getting quite the answers you want, yes, absolutely do the prompt engineering if you’d like. Mostly what we’re thinking is that that is an optional thing rather than a mandatory thing, that you don’t have to prompt engineer, but you certainly can if you’d like.
Marlene Gebauer (04:54.254)
All right, so the expert prompters aren’t dead quite yet.
Damien Riehl (04:57.238)
100%. Yeah, much like Boolean searches from years gone by, you can be able to use those. Exactly. It’s still beautiful. But yeah, I think that the more we’re doing our jobs, and the weird place that we’re in with large language models is that the large language models can actually improve prompts. That is, you’d be able to say to the large language model, take this prompt and maybe make it better for GPT -5 or GPT -6 or Lama 3, whenever it comes out.
Marlene Gebauer (05:00.654)
Still kicking.
Damien Riehl (05:24.086)
So we’re in a weird space that the human ability to prompt is maybe decreasing in necessity as the large language models themselves learn how to improve prompts.
Marlene Gebauer (05:35.278)
Um, so how much time do you think this actually saves? I mean, you basically like distill that entire workflow and, and I’m just curious, like, how much time do you think?
Damien Riehl (05:48.854)
A lot of time? Yeah. When I litigated for a big law firm, the amount of time that I think it would have taken for an associate to do what I just showed on my screen, I would estimate maybe 40 or 50 hours. That would be a full week’s worth of work to be able to develop all the analyses, to be able to do the research. So maybe 40 or 50 hours, and it takes just minutes, not hours. And so the real question is,
Marlene Gebauer (05:49.966)
It’s kind of crazy. It’s like, it’s like, boom, it’s done. And, and.
Damien Riehl (06:17.558)
What is the effect on the market? So we sell to law firms, obviously. We also sell to their clients, corporations. So we have Fortune 50 companies that are using this and are saying, wow, this is amazing. What we’re going to do is two things. Thing number one is we’re going to take all the complaints that were filed against us in the last year, run it through the system, and then compare it with what our law firms did and say, what did they do on top of it, if anything? If the answer is nothing, they didn’t do anything in addition, that’s going to say something.
And then they said going forward, what we’re going to do is we’re going to take every new complaint that we received and we’re going to run it through this before we send it to the law firm. And then we’re going to send this output to the law firm. And then we’re going to say, what can you do on top of this? Because really that’s all we’re going to pay you for. All we’re going to pay you for is the ability to be able to, right, we’re not going to pay you to recreate the wheel. So I think we think that this provides law firms that are forward thinking and the clients that are forward thinking with their law firms to be able to do a lot more work.
Marlene Gebauer (07:03.15)
the extra.
Damien Riehl (07:13.622)
and more efficiently and to be able to do more work. And maybe this ushers in going from the hourly model to the flat fee model, where maybe you’d be able to say if it would have taken 50 hours in the hourly model, maybe charge a flat fee of the equivalent of maybe 20 hours or something. And then it will maybe take only three hours to do the thing instead. And then you can be able to make all of that extra in profit. That might be one of the future worlds that we’re looking at.
Marlene Gebauer (07:32.238)
Yeah.
Marlene Gebauer (07:39.31)
pricing people are going to have their hands full. So let’s talk about this exciting new hire that you have.
Damien Riehl (07:41.302)
Indeed.
Damien Riehl (07:47.382)
Yes, Dan Hoadley out of the UK, formerly of Mishcon de Reya LLP, is now going to be starting with us as of I think today. So he is, for those who know the legal tax circuit, he has a brilliant, brilliant mind. He’s really not only a brilliant mind and doing really some cutting edge work for Mishcon de Reya LLP and was working with an amazing team with them, but he’s also a really wonderful, nice human being. If you’ve ever met Daniel, he is really…
a good person and we really couldn’t be more thrilled that he’s joining us. He’s going to be leading up our research and development team where he’s going to be thinking deeply about how we can take our billion legal documents and be able to do data science on those billion legal documents and be able to extract what’s needed from those and then be able to run large language models across them. So we think that Dan is really an exciting, exciting addition to our team.
marlene (08:41.038)
That’s terrific. So do you have any kind of up and coming projects for him that are like on the roadmap?
Damien Riehl (08:48.534)
We do. So the Skunk Works, which Ed Walters, yeah, so Ed Walters and Angel and Robin on our Skunk Works team, we have lots, we have a list, our arms length long. And so it’s really, the next is gonna be an embarrassment of riches. Of that long, long list, what are we gonna do first? And that’s what Dan gets to figure out. So yeah, so stay tuned. There are gonna be some really good things coming out of Dan’s R &D team.
marlene (08:50.894)
that you can talk about.
marlene (09:12.814)
That is terrific. Well, welcome, Dan, and congratulations, Villex, and Damien, thank you for joining us.
Damien Riehl (09:19.382)
Thank you so much for having me.