The legal system is complex, hard to understand, expensive to navigate, and ripe for disruption.
In the future, we will still need lawyers to help us understand the law, but it look like we are going to need far fewer of them than we have today.
Nozo Tsunoda is an attorney who walked away from a promising legal career to start LegalOn, an AI startup focused on making the practice of law more efficient, transparent, and easy to navigate.
We talk about why corporate legal departments are the early adopters, but why AI technology is forcing its way even into the most traditional law firms, and how it might someday be used by consumers as well.
It’s a great conversation, and I think you’ll enjoy it.
- Why it’s hard to sell AI technology to law firms
- How AI is starting to change the way law firms compete
- Why Nozo left the law to start a legal startup
- The contract review workflow and why it’s perfect for AI disruption
- How many lawyers will AI replace in the next five years?
- Differences in how US and Japanese staff view working from home
- A $100M investment in US market entry
- Differences between Japanese and American legal systems
- Can today’s AI understand contracts better than a junior associate?
- The big changes AI will force on the legal industry
- The need for more immigration in Japan
Links from the Founders
- Everything you wanted to know LegalOn
- Their US website
- Learn about LegalOn’s Products
- Read about LegalOn’s US market expansion
Welcome to Disrupting Japan. Straight Talk from Japan’s most successful entrepreneurs.
I’m Tim Romero and thanks for joining me.
Of all the industries that are going to be impacted by artificial intelligence, the legal profession is going to be one of the most profoundly transformed.
And today we sit down and talk with Nozo Tsunoda, a licensed attorney and the founder of LegalOn a rapidly growing startup using AI to review and manage contracts. And while their initial clients have been mostly corporate legal departments, Nozo explains how AI is beginning to force changes to the behavior of even traditional legal firms.
Now if you’re in the US you might not have heard of LegalOn yet, but you’ll be hearing a lot about them soon. Nozo and the team recently raised over a hundred million in large part to fuel their recent US market entry.
Nozo and I talk about the challenges of selling increased efficiency to lawyers who bill by the hour LegalOn’s US expansion plans, and how AI is going to change the entire legal industry.
But, you know, Nozo tells that story much better than I can. So, let’s get right to the interview.
Tim: So, I’m sitting here with Nozo Tsunoda of LegalOn, who’s using artificial intelligence to simplify and improve contract review and management. So, thanks for sitting down with us.
Nozo: Thank you.
Tim: Well, listen, I gave a really brief description of what LegalOn does, but I’m sure you can explain it much better than I did. So, what does LegalOn do?
Nozo: LegalOn technology is legal tech companies. I found it seven years ago, and now we have three solutions. And globally we have four solutions and globally we have 3,700 customers.
Tim: Well, but to get to the basics for some listeners who might not understand anything about the legal process, so what is the service that LegalOn provides?
Nozo: For contract area we have three product for pre-ex execution process of contract drafting or review. And second product is for contract management.
Tim: So, contracts is a very, very broad subject. So, LegalOn focus is mostly on things like NDAs and purchase agreements and things like that.
Nozo: Yes, of course we can support NDA, purchase agreement or service agreement, but we can review 50 types of contracts for the market.
Tim: Tell me a bit about your customers. So, you said 3,700 customers, which is fantastic. Who are they? What kind of customers are they? Who are you selling to?
Nozo: Enterprise corporations or midsize corporations.
Tim: Okay. But it’s mostly selling to corporations, not to law firms.
Nozo: Both. So, we have 500 law firm customers.
Tim: Now that is interesting because in one of the many startup projects I’ve started over the years, one of them was actually a contract lifecycle management company. And what we found was that it was incredibly hard to sell to law firms because they bill by the hour. They don’t want to improve efficiency, they don’t want to do things any faster. So, how are you selling to law firms?
Nozo: So, law firm can use our product as their tools or their weapon. They can improve their productivity or quality of their product…
Tim: So, most of your customers are enterprise or corporate customers. And there it makes sense because in-house lawyers, they want them to be as productive as possible. If you can get the work done with three in-house attorneys, that’s much, much better than 10 in-house attorneys. But if you’re a law firm, you want to have 10 attorneys billing the project.
Nozo: So, I think they’re competition, law firm need to win new clients by their service levels or service qualities. If they use our product, they can support more customers.
Tim: That’s interesting. Actually later on I really want to talk about how the legal industry is changing, but from your perspective, the law firms are responding to this kind of pressure and competition.
Nozo: Yeah. In Japan or maybe globally law firm tried to get new customers. Because then they compete by their services. So 10 years ago or 20 years ago, lawyers didn’t have such kinds of ideas, but now law firm need to think about their service.
Tim: That’s a really positive development. That’s good for everybody. Well, actually, before we talk about the product, let’s talk a bit about you. So, you’re actually a licensed attorney in Japan, right?
Tim: Now I cannot imagine two career paths more different than being an attorney and being a startup founder. So, what made you want to make that change?
Nozo: So, then I became attorney I didn’t think I will be product founder. But 10 years ago people think about deep running or math running and I found it legal with co-founders and we discuss about if we can leverage machine learning or deep planning to legal areas, we can change legal practice or improve.
Tim: So, he was your colleague at Mori, Hamada and Matsumoto law firm. Talk to me a little about that decision. Because it just seems like such a big — so I talk to a lot of university students who become startup founders, but going from an attorney to a founder just seems like such a big jump. So, what made the two of you decide to do this?
Nozo: We thought this change of technology is a kind of must, must come in the future. So, if we didn’t try legal practice, we be changed by another person.
Tim: Yeah. Someone else is going to do it.
Nozo: If so let’s try. It’ll be exciting.
Tim: No, that makes perfect sense. But at the same time, or close to the same time, you also founded the ZeLo Law firm. What was the connection?
Nozo: So, now I focus on management of LegalOn and co-founder, Ogasawara, we founded both LegalOn and law firm ZeLo. And at first phase I worked for ZeLo as an attorney to earn money.
Tim: So, was this your backup plan?
Nozo: At first phase we had a plan. So, legal developed technologies, then ZeLo leverage technology to legal services. So, we tried to create kinds of new legal services with technologies.
Tim: So, ZeLo still exists. What’s relationship today? So, I mean, clearly your LegalOn is more than a full-time job.
Nozo: So, Masa manage ZeLo and ZeLo is expanding too. Now it has 100 people.
Tim: Do the companies just share a history or is there a closer relationship? So for example, a lot of your advantage is that the suggested language is reviewed by attorneys. So, is that done by LegalOn attorneys or is that done by ZeLo attorneys or?
Nozo: Both. Legal attorneys make directions about contents, but we order ZeLo attorney to create new content source and they’ll use our product. This kind of collaboration is now…
Tim: That’s interesting. So, it’s still very close relationship. Let’s talk a bit more about the product. So, what happens during contract review, walk me through the workflow of how an attorney using LegalOn will review a contract and why it’s better that way.
Nozo: To review contract attorney need to find this point but it’s really difficult to realize if there are nothing. So, if there are some this point it’s written, we can find it. But if there are no clause, we should add some clauses.
Tim: So, the AI will go through and highlight the risk points, whether it’s wording that might not be quite correct, or clauses that might be missing and should be added. What does it show the attorney who’s using it? Does it say you should consider putting in this text or this word is dangerous or what kind of suggestions does it make?
Nozo: So, our product can show missing clauses or less clauses in few seconds, 10 or 20 points. Then attorneys they consider about that point and then change document.
Tim: So, when an attorney has to review a 40 page contract, LegalOn will highlight, these are the 10 points you should double check and the rest is okay.
Nozo: Yeah. It’s difficult to say rest is okay but our product can find 10 point…
Tim: For special attention.
Nozo: Yes. The attorney need to check, then the attorney can reduce time.
Tim: So, it’s more of almost like best practices type of advice. These are the kinds of clauses and kinds of wording you should think about using in this case.
Tim: So, the reference text that it displays, the reference text is not generated by AI, it’s generated by real attorneys. So, within LegalOn what’s the breakdown between what percent of the company is legal, what percent is like sales, what percent is engineering?
Nozo: 40% for development and 50% for marketing and sales and 10% is corporate.
Tim: And are the attorneys considered part of development?
Tim: So within development, what’s the percentage of engineers and percentage of attorneys?
Nozo: 80% engineers.
Tim: Well, that’s still quite a few attorneys. I mean, contracts seem like almost the perfect application for artificial intelligence. It’s very specialized language. It’s very formalized and a contract is a lot of language and lawyers are expensive. So, it’s almost perfect. So, you mentioned that you support about 50 types of contracts. So in the future, like five years from now, 10 years from now, what percent of corporate legal work do you think AI will be able to do?
Nozo: Difficult question. So, I think AI cannot replace human but strongly support, I don’t know about 50% or 70% AI can support.
Tim: So, in the future, for example, if we have a corporate legal team who is 50 people today, that same team doing the same workload five years from now with AI support will be like 15 people or 20 people in the future.
Nozo: In the future.
Tim: Yeah, that makes sense. So, what’s LegalOn’s engagement model when you get a new customer? Is it a pure simple SaaS model or do you have to do a lot of training and customization for each firm’s legal needs?
Nozo: For the SaaS model, yeah.
Tim: So, I imagine all companies have their standard, like every company’s NDA is almost the same but different. And every company’s employment agreements and it’s like every company is different. How do you work with that?
Nozo: So, if some company receive contract from third party, they need to check at that phase. Our product can strongly support cyber party contract and if our customers use their template, we have some function to support such situations.
Tim: Okay. So, the customers can put in their own wording for this situation.
Nozo: So, our customers can use some functions depending on situations.
Tim: So, they can customize it a little bit themselves. So in 2020, in response to COVID, the Japanese government, I mean they announced all kinds of changes to legal guidance and in some case the law itself to support digital transformation. Things like not requiring hankos and trying to push really hard to get everyone to digitize. Has that helped your business?
Nozo: I think so. Before COVID digitization or digital transformation was for some early adopter companies, but after COVID many company realized this digitization is needed.
Tim: So, during COVID companies were adopting digital transformation because they had to, there was no other choice. So, we’ve been out of COVID, let’s say about a year, six months a year. Do you see enterprise customers going back to the old way or do you see them more staying with more digital tools?
Nozo: Talking about this digital, many companies will not go back. They try to use digital tools or they think digitization is really important. And now Japanese government push…
Tim: They’re pushing it harder now.
Nozo: Yeah. But working from office or working from home, many companies back to working from office.
Tim: I did not think that was going to last.
Nozo: But this digitization is not.
Tim: It’s interesting the difference of opinion in working from home in Japan and working from home in the US. In the US most people wanted to work from home. Like if they could just work from home every day, they’d be very happy. In Japan, I get the impression most people want the option to work from home once in a while, but generally want to come to the office.
Nozo: It depends on size of their house.
Tim: Yeah, that’s true.
Nozo: Many house does not have working space.
Tim: Yeah. Small Tokyo apartments.
Nozo: So, if there are some kids, it’s hard to concentrate on their jobs.
Tim: Yeah, that’s true. The family life is driving a lot of people back to the office. Well, let’s talk a bit about your US market expansion. So, you recently raised about 101 million from SoftBank, Goldman, and Sequoia. And a lot of capital that’s targeted on the US market entry.
Nozo: So, I respect legal practice of the United States. In legal areas United States did globally, so many Japanese corporate lawyers go LLM. LLM means kind law school for lawyers. So, many corporate lawyers go United States and then they run practice and then they spread the new practice in Japan. In bigger areas, practice of United States is progressing. So, for contract or legal space, the United States is mostly in product market in the world.
Tim: Now you said something I thought was really interesting. So, when you were talking about the US as a leader, I understand that US contract law is very important all over the world. But do Japanese lawyers view the US legal system as like this is a really good system or…?
Nozo: It’s difficult to judge which legal system is better, but the practice of the United States is really important.
Tim: It’s very different. Japan and the US. So I mean, as someone has to deal with contracts in both countries, if we do a same contract, whether it’s an employment contract or an NDA or a stock purchase, in Japan it’ll be like a three page contract. In the US it’ll be like 60 pages. Why is that?
Nozo: I think it depends on the litigation system. So, litigation risk is really high in the United States. So to try new transactions, it is important for both parties to write in contracts.
Tim: One thing I’ve heard about lawsuits in Japan is that the Japan legal system makes it very difficult to sue anybody. Usually both parties lose in Japan
Nozo: Yes. So, in Japan court system is really severe to plaintiff.
Tim: So, in Japan is a lot of the litigation handled just through private negotiations and that makes sense. So, in the US its people need all the extra language to protect themselves.
Nozo: By contract.
Tim: But I guess that would make it even a better market for LegalOn.
Nozo: I think in the United States, the importance of contract is higher than Japan. So, there I think there are opportunities for technologies like LegalOn review.
Tim: Well, yes. And also, I mean a customer will pay much more when they have to review 60 page contracts than we have to review three page contracts. Because the law in the US Japan is so different, did you have to change the product? Did you have to change like some of the core functions within LegalOn?
Nozo: So, we need to make product rules that fit the market of the United States because legal practices are different, but maybe we can leverage the core functions of Japan product.
Tim: Getting back to kind of general AI concepts again. So, now I mean attorneys need to be licensed to practice in a particular country or in the US in a particular state. Do you think as AI is increasingly used, that they’ll be similar requirements on AI systems where Texas courts will require certification for this AI and…?
Nozo: Our system is for a kind of tools for legal professionals.
Tim: But the technology AI is moving so quickly. That even though today it’s kind of a background tool. Putting aside that one stupid lawyer who like submitted chatGPT to courts.
Nozo: I’m really impressed too.
Tim: How did he pass the bar? But putting that aside, even if today it’s all back office and support, in the years to come it’s going to move closer and closer to the actual courts. So, do you think it’ll come a time where it will have those licensing?
Nozo: I think there are some points. One point is how technology will progress and now performance of LLM is excellence really surprising, but there are some kinds of problems. So, it cannot support the accuracy of the content. So in the future, I don’t know if we can solve that kind of problem.
Tim: So, even though today with LegalOn, we still have an attorney making the final decision. We just have the AI analyzing and pointing out potential problems. But as we were talking before, like it’s such a good application for AI contracts or like patent law maybe even better. Do you think we’re at the point yet where AI can understand a contract as well or better than a new junior partner?
Nozo: It’s difficult question. So, from my understanding LLM is system to create natural range. So, it seems like it has knowledge but it’s difficult to know real mechanism of neural network. So, junior associate can think by their brains and their knowledge.
Tim: But I mean, you don’t trust a junior associate to make the final decision either. So, I mean it’s like…
Nozo: So, now lawyers who use AI or lawyers who use junior associate but he need to decide, may need to make final decisions. So, in the future if we can delegate final decision to AI, so it depends on the level of AI. So, if AI can…
Tim: Well, I mean that’s what I think is so interesting here is where we’re going to be in five years or 10 years in the future. Because so if we say that, okay, today’s AI on the level of a junior associate, plus or minus. They’ll pull out the same type of questions and great, but as you say, it’s going to get better and better and move higher and higher. Well, do you think we’ll get to a point where AI is supporting or actively running even the negotiation where you have this company’s AI is negotiating with this counterparty’s AI to…?
Nozo: It’s interesting. So, if one company delegate AI to negotiate and then one company delegate to AI to negotiate. It’s interesting. I cannot imagine what’s happened.
Tim: Well, I don’t know. I mean I think it could be really useful because the AIs know the internal company goals. They know what they can give and what their different values are. So, I don’t know, maybe it would be a very efficient way to negotiate. But do you see AI being used like it’s not just the review process, but in the negotiation and drafting and the whole contracting process.
Nozo: I think it can be useful because if AI negotiate from both parties, I think there are potential, they can create great solutions.
Tim: Better than humans.
Nozo: Yeah. If people couldn’t find great solution, but if AI did that GPT can create.
Tim: Well, I think we’re still at a very, very early stage, but I think there’s another aspect that people don’t talk about much. I think this is going to have to change the structure of the legal industry. For example, before Excel or actually before VisiCalc, if you’re old enough to remember that before spreadsheets, accounting companies employed lots and lots of bookkeepers. People’s whose job it was just to add up columns in both directions, make sure they agreed and the career path was, you’d spend a few years as a bookkeeper doing this, and then you’d move up and become an accountant and move up in accounting more strategic. But when spreadsheets came out, this whole job just disappeared. And it seems to me the kind of technology you’re introducing like in legal firms, junior associates spend a few years doing this contract review and highlighting potential risks and unusual things and then they move up and it seems to me like that’s going away. So, how do you think this is going to change the legal industry?
Nozo: So, I can understand your point because I worked as one of the associates. And I spent a lot of time to contract works or document checks or some portion of associate works will be changed to technologies. But I think not everything.
Tim: I mean, continuing the analogy with spreadsheets, it’s not going to be just lawyers using these tools. As costs come down and quality goes up I would imagine consumers would also be using these tools for a lot of simple things rather than going to lawyers.
Nozo: Talking about futures contract review process can be separate to some processes, check document, then find risk or find missing. And there are some candidate of collect and then decide and negotiate decision making process is difficult to replace by AI or technologies.
Tim: But what about when we’re talking about standard contracts, as a consumer if I’m signing a new lease or want to make a will or signing an NDA or an employment agreement or one of these simple things that today we have to go to a lawyer, doesn’t that seem like something an AI would be doing for consumer advice?
Nozo: So, if needs of consumers is to find this point or something. So, AI can support.
Tim: So, something like, hey this employment contract has a non-compete that’s not good for you…
Nozo: AI can do that.
Tim: So, I see this technology as being really, really disruptive in the true sense of the word to the legal industry. I think the traditional work to junior associates is going to be going away. It seems to me once this technology becomes available to consumers, a lot of the very steady, predictable work that lawyers do.
Nozo: So, it depends on what consumers want. So, if consumers need comment, AI can support, but if consumers want advice, it will repeat to lawyers.
Tim: Well, I guess legally it couldn’t give legal advice. Legally speaking. Lawyers are protected that way. But how soon do you think we’ll see this kind of technology being used by consumers?
Nozo: Our customers are corporations or law farm. So, we don’t serve our product to consumers because it is difficult for consumers to understand commentary show.
Tim: That makes sense. You’re suggesting proper legal boiler plate that a consumer might not…
Nozo: To fix contracts, people need to find risks, then this make decisions then refer to contract and then fix. So, it’s difficult to make decisions.
Tim: So, as a consumer product, it’s not a complete product. It can highlight the risk but the consumer won’t be able to act on that information.
Nozo: Yes. And I think LLM has same problems. So, if as long as more they create contract, consumers cannot just, it’s good or not. So, now AI’s good tools for expert.
Tim: Oh, that makes sense. Kind of disappointing. I’m really looking forward to having an AI lawyer. It would save me so much money.
Nozo: So, if some companies create perfect AI we can believe it’s cool totally.
Tim: But no, I think you’re right. That final step, you have to get a lawyer involved no matter what at that last step. Well listen Nozo, before I let you go, I want to ask you what I call my magic wand question. And that is, if I gave you a magic wand and I said you could change one thing about Japan. Anything at all, the education system, the way people think about risk, how quickly people adopt new technology. I mean anything at all to make it better for startups and innovation in Japan, what would you change for Japan?
Nozo: We need more people and global peoples. So, because now populations of Japan is reducing and if this situation continue, Japan will be shrink. We need more populations and we need to change to more globally.
Tim: So, do you think the best kind of increasing population would just be increasing Japanese birth rate or more foreigners coming to Japan?
Nozo: I think both. Both.
Tim: Both. So, I’ve been in Japan for 30 years and it’s very interesting to look at how the social attitudes of foreigners coming to Japan has changed in that time. So, Japan used to be very conservative about this kind of thing. And it’s really changed. So, in terms of like innovation, what’s the big advantage of having more foreigners coming into Japan?
Nozo: Great impact is domestic people can connect to another countries. I think many Japanese cannot imagine they can make business with global partners.
Tim: Also just having more foreigners in Japan makes Japanese more globally minded. Just the everyday interaction.
Nozo: So, now we are try to go to United States, but it depends on JPs. So, JP is key persons of our global expressions. So, before I met JP I couldn’t imagine I can corroborate with you or I can go to United States.
Tim: So, in LegalOn, do you have a lot of non-Japanese working?
Nozo: Talking about persons, it’s not enough. But now increasing I think.
Tim: Well, I’m not sure about the changing of the birth rates, but it does seem like Japan is becoming much more welcoming to non-Japanese. There’s more and more foreigners coming to Japan, particularly engineers and designers and so that seems to be happening now.
Nozo: Good news. I think Japan is good to live. So, now Tokyo is changing so easy. There are many English guides.
Tim: Oh, it’s so much easier than it used to be. Well, listen Nozo, thank you so much for sitting down with me. I really appreciate it.
Nozo: Thank you very much.
And we’re back.
After that interview, I spent a lot of time thinking about the potential impact of legal AI. And a legal AI is clearly going to drive huge changes in business and technology, but I think it’s going to have an even larger social impact.
You know, it’s strange, people always overemphasize the impact of technological changes, but we tend to downplay the impact of social change. And that’s not right.
For example, I’ve heard so many people make the claim that if you took the average American from a hundred years ago and brought him into the present, he would be dumbfounded by smartphones and the internet and dazzled by modern technology and would be unable to function in modern society.
And that’s just wrong.
I mean, these things are just tools, handheld communicators, knowledge at your fingertips, space exploration. I mean, these were already established parts of science fiction at the time. Our time traveling friend would undoubtedly be impressed and excited, but after a few weeks he’d be posting selfies on Facebook and doom scrolling on Twitter just like the rest of us.
The social changes he would see however, would be shocking. The fact that America’s had a black president and openly gay and lesbian Congress members, the degree to which social violence has become unacceptable. The extreme value we place on human life and worker safety, the fact that many of the most influential people in the world are women. Now those things are probably touching on some deeply held beliefs about how the world works and the way things are supposed to be. And there’s a good chance our time traveler would never really be able to understand or accept it.
So, with that in mind, what social changes will legal AI usher in?
I think we will see AI license to give legal advice in limited contexts. For example, AI will advise consumers on the fairness of leases, rental agreements, employment contracts, and help prepare wills and trusts. These AI legal services might even be available for free, something the government provides as a natural way for its citizens to interact with the complex legal code and to help ensure that contracts are fair and understood and relatively standard.
But AI will be even more transformative in contract negotiation. Unlike humans, AI can be programmed and certified to fairly represent both parties in a negotiation. It can keep confidential information and strategic objectives compartmentalized and secret.
Now such a system could reliably, quickly and inexpensively produce a negotiated agreement optimal for both parties or advise that there’s no deal to be reached. And once an AI can be programmed and certified to be an honest broker, to never cheat, mislead, or favor either party, they’ll become the default. I mean, why would you want to do business with anyone who refuses to negotiate within a fair system?
The whole adversarial nature of civil law would be transformed into a cooperative one. A hundred years from now people will consider today’s adversarial legal system as barbaric and wasteful and shake their heads as they try to imagine why any honest person would do business that way.
Oh, we’ll still need lawyers in the future, of course, but we are going to need a lot fewer of them.
If you want to talk more about how AI is going to change the legal industry and change the world, Nozo who, and I would love to hear from you. So, come by disruptingjapan.com/show207 and let’s talk about it. And hey, if you enjoy disrupting Japan, share a link online or just tell people about it. Disrupting Japan is free forever and letting people know about it is the absolute best way you can support the podcast.
But most of all, thanks for listening and thank you for letting people interested in Japanese startups know about the show.
I’m Tim Romero and thanks for listening to Disrupting Japan.