Evisort CEO On How Generative AI Will Change Legal Sector
January 20, 2023
Evisort CEO Jerry Ting talks with Law360 Pulse about Evisort's generative AI and how the technology will change the legal profession.
Evisort SEO On How Generative AI Will Change Legal Sector
Earlier this week, contract management software developer Evisort said that the company is rebranding to align with its December release of a generative artificial intelligence tool for contract drafting.
The company's rebranding includes new product design, a revamped website and more resources for customers about AI and how to use the technology. Generative AI can produce original images or writing, like a contract, rather than simply analyzing existing data.
Evisort CEO and co-founder Jerry Ting told Law360 in an interview Wednesday that the company decided to rebrand because it wants the brand to feel more human and be less tech centric.
The rebranding "coincides with the fact that AI is going much more mainstream, and we're having conversations right now with thousands of different people, on a per-month basis, that have never used AI before, and so we want to make sure we speak to them correctly," Ting said.
Here, Ting talks with Law360 Pulse about his company's generative AI and how the technology will change the legal profession. This interview has been edited for length and clarity.
How long did it take the company to develop the generative AI that it released in December?
It took us over two years. Just to give a bit of background on how we're using it. Our users, which a lot of them are lawyers, really [use it] in the redlining process. So you're sending a contract out, somebody redlines it [and] sends it back. That becomes a very manual process because inherently every single redline is third-party data. It's generated by somebody else that's not in your company. And so you have to read every line and make sure you agree with it. Then, figure out what to write back.
What we were doing is we had to solve that entire problem, and that was an over two-year process. The first step of that was, "Hey, he sent the redline back. What is it saying? Are they negotiating this clause? Are they negotiating this business term?" And so the first part of that was to get the AI to understand that, and then, once we did that training the AI models so that it can actually write the response back. A lot of times people talk about generative AI just drawing pictures and writing marketing blogs. Ours is much more deeply applied because it actually needs to understand the meaning and respond to the context of that meaning.
What was the biggest challenge of creating generative AI for contracts?
It's getting the AI to write in a way that a lawyer would look at and say, "Hey, that's like 80 to 90% the way I would do it," and then, edit it to make sure that it's complying with their drafting style. That took a lot of time.
Also, [getting the AI to understand] what are the playbooks and the policies that a company wants to apply to a contract, because that can vary from company to company, from industry to industry. It's very nuanced, and it also changes based on the type of contract that you're negotiating. It's very different to negotiate a five-page confidentiality agreement versus a 40-page services agreement. So there's a lot of nuances in diversity in that data.
What exactly can your company's generative AI do? You mentioned the redlining process. Is there anything else that it can do?
It can actually help you write the first contract too. The way it works right now is if a salesperson needs a contract to be created by a lawyer, they can fill out a form directly inside of Salesforce, and then, based on that form, using generative AI, we can actually write the contract, and a lawyer just has to check it. It makes that drafting process a lot easier. Some examples are sales proposals, statements of work and purchase orders.
What does the company want its generative AI to do in the future?
I think what I'm describing here is pretty transformative already, but I think the next chapter is getting very industry specific. If you think about how lawyers are trained, I was trained to be a [mergers and acquisition] lawyer. My law school classmate was trained to be a banking lawyer. Being able to get [generative AI] to be industry specific where it can negotiate a loan agreement that right now is still a frontier item.
It can negotiate sales contracts. It can negotiate vendor contracts. But to negotiate a $200 million loan, that's where it's still just 100% manual, and law firms charge a lot by the hour. I think getting into these more complicated agreements that have more specialty applications will be the next chapter here. And also multiple languages. We just wrote out a software that can read over a hundred different languages, and that's new.
Why do you think there's so much hype around generative AI right now?
I think the reason for that is now the technology has matured enough. There's been training data [and] enough computing power where even my parents can use it in a small business. My mom, who speaks English as a second language, actually asked me, "Hey, how do you use generative AI to edit contracts?" Everyone's thinking about it. It's because the technology has come far enough where it could be applied rather cheaply. Five years ago to use AI for a company, for a dollar of AI, you would have to spend $2 to $3 with a consulting firm to actually get it to work. Today, with our AI, it works out of the box. It works for small companies and large companies, works day one [and you] don't have to set it up with a lot of services.
What feedback have you gotten from clients that have tried your generative AI?
They think it's kind of magical. [Lawyers are] taught, trained and conditioned to think what [they] do is just very special, but then when you actually go practice law, there's a lot of very manual things. So when our customers started using it, and we have a couple of design partners that we actually built this with, they said, "Wow, I wish I had this 20 years ago because I could have got my thirties back." Now they're figuring out how to just roll it out more broadly. When you think about a new technology like AI, the first step is understanding what it is. What can it do? What can it not do? The second step is: How do I put that into my business process? A lot of our clients are going through that second phase now of "Hey, we know it works. We've tested it thoroughly. How do we get 400 people to use it?"
When you think about a new technology like AI, the first step is understanding what it is. What can it do? What can it not do? The second step is: How do I put that into my business process? A lot of our clients are going through that second phase now of "Hey, we know it works. We've tested it thoroughly. How do we get 400 people to use it?"
Is there any legal guidance out there for attorneys about what they can use generative AI for and what they can't use it for?
It is almost a completely unregulated space. It's actually one of the areas where, for me, I would actually like more regulation and more definition. As a lawyer under bar licensing, do you have to review it? Do you not have to review it? If we have clarity, it would actually make it easier for people to adopt because they can say, "Hey, for these things, like e-discovery, I don't need to look at. But for these things in drafting, if I'm representing a client, I do need to look at." That would actually help us design software that is actually better aligned with our users. But right now, because the technology is just so new, I think it's figuring out what can it do and how do we use it.
The way that I would recommend lawyers use this is as a tool to help them with what they're doing. For me as a lawyer, I wouldn't feel comfortable saying, "Hey, the AI did it," and then, send it over. You got to take a look at that. So for me, it removes the 80% of the boring, repetitive and tedious, but it doesn't take the lawyer out of the picture. Just like when you're driving a Tesla, the driver is in the seat [and] they're still aware. That's where we are right now as an industry. I suspect it's going to be like that for the next decade or so. Even if the technology works, it's not about the technology. It's about fulfilling your obligation as a professional.
How do you think generative AI will change the legal profession?
I think it's going to make the legal profession more strategic and more fun. I went to law school to actually become a lawyer. I'm not an engineer. I don't code. The reason why I left the law is a lot of the stuff was very manual [and] tedious. Go through a 30-page contract, look for a couple of specific words, if they're there, do this red line, and that's not why I went to law school. I went to law school so I can help protect against risk and accelerate businesses and make sure things are done correctly. That's the strategic parts of being a lawyer. But most of the time, for lawyers inside of companies, they're reading contracts and supporting sales teams. It's important, but it's very repetitive and should be partially automated.
Find out how
can help your team
Volutpat, id dignissim ornare rutrum. Amet urna diam sit praesent posuere netus. Non.