The Intersection of AI and Contract Creation: Insights from Pavel Kovalevich

Hi, Harshit. It’s great to be here. Super happy about the introduction. Yeah, looking forward to our chat.

I think to talk about this, We might go back a little bit when we started, and that was around 2016. I think when we started we didn’t have any customers, but we had some ideas about where the macro shift was happening. One of the things that we have seen was that there were not a lot of good products out there for legal teams. That’s number one. For those that were, they were focused mostly on the legal part of the team. Our insight was that a contract is essentially the lifeblood of the business. That means in practical terms that a commercial part of the business will not exist without a contract. That’s a commercial business 101. That means that it’s not just legal that participate in those contracts, it’s everything. It might be the sales team, it might be the CFO, operation, you name it. It’s probably one of the not-so-many business areas that touches on everything in a business. When we started, we wanted to have a very human-centered approach to creating the software. What that means is not just focusing on a particular role or in a particular functional use case, but focusing on the people that are involved.

That means focusing not only on legal but focusing on all of the people who interact with the contract, all of the people who interact with the negotiations. That was number one. From day zero, we had a very human-centered approach to both designing, building, figuring out what to build, interacting with customers, and even hiring. So across the board, we wanted to have a very human element in sometimes the very mechanical analytical approach to building software. So the design was very important. The second point was data. So one of the interesting things that we found for ourselves, and we thought that there’s quite a lot of value in was the data aspect. And that data aspect was the foundation where we said, if we can help people organize and gather their data, then that data represents value. Usually, the classical way is that if you have a contract, you print it out, you sign it, you fax it, whatever, and then it lives in your drawer. Maybe once a year, you take it out, and a bookkeeper takes it out. But that’s pretty much it. That value stays locked inside the drawer and no one has access to that value.

We’ve seen value both in the data aspect of the whole commercial contract life cycle and the design aspect. That’s where we started. Sometimes it was not the most optimal way to do things, but that was part of the value proposition that resonated with us, with the team, and honestly, with our customers.

Yeah, of course. The AI aspect is very top of mind for a lot of people. Everyone who was not living under a rock is just basically thinking about it, either building something or integrating it into their products or stuff like that. I started by talking about the data aspect, which was the original founding vision for Juro, where we said we could build a chatbot that could analyze your contract. That was 2016. For people who don’t remember, 2016 was the year of chatbots. That was the year of chatbots. I think it was like, Slack was pretty high back then. They even had their fund for chat-based apps or something like that. I think that was part of it. The data aspect was part of it. We tried to build a legal system in the form of a chatbot, and we failed miserably. After one and a half years, we said, Okay, we solved some of the things that we wanted to solve, but honestly, the tech is not there. That was after a year and a half of working on an AI chatbot, we said, We’re going to shut it down.

We’re going to carve out those pieces that work, integrate them into the product, and wait until the tech is there. I think the last couple of years, for everyone who is in the industry, it’s clear that the tech is there. Maybe not fully. Maybe we still have a little bit to go, but it’s already there. I think the first thing that we did was we came back to that original vision of how can we enhance the experience of the user that interacts with a contract. For us, it’s the three distinct phases because we’re an empty-end workflow. The three distinct phases are the pre-signature, post-signature, and negotiation flow. Essentially the pre-signature, for example, will help people by identifying how should the pre-signature flow go. Where are the things that need approval, for example? What are the drafting pieces that need to come together? All of this is done through an interface that is as close as possible to a one-click solution. I think a lot of them, especially when this software scales, there’s a key aspect, the speed aspect, and the risk aspect. For the speed aspect, it’s Since we know quite a lot about the workflow of our customers, one note to say that it doesn’t work in absolutely all of the cases.

I think it works only because we know for sure what the workflow of our customers is. That allows us to have a set of assumptions encoded into the product. Another piece is figuring out, for example, what piece of the document represents a risk from a company standpoint. Do that instead of a lawyer or legal assistant or an external firm reviewing the document, we can do that automatically because of AI. That’s the short one.

I think it’s a very multi-layered question, I think. There’s not one piece of feedback that we got.

I think that’s a good way to start because I think for Juro, there’s no specific, for example, there’s very little value in specific feature feedback. We always start digging into how this relates to the problems that our users are having. There are a lot of questions that get asked along the way. I think all of this is because working with really understanding customers, working with their feedback, working with their comps, it’s really in the DNA. When we started and when we started with the human-centered design hat on, I remember the first thing that I did, this was before we had a product or customers or anything, was I created the Google Sheet. It was funny because at first, this was just a way to record conversations, but very fast, the whole team was using that Google Sheet to record all of the feedback that they found interesting and remarkable from any conversation with anyone that they had, whether it’s an investor, whether it’s a potential customer, because we didn’t have any customers back then, like prospect, whether it’s with competition, whether it’s with a new hire. Anything remarkable was reported there, and it was a very long list of stuff.

It was unfiltered. I think that made it very easy for people to record because this is just per conversation, just record a line or two, of what you remember, especially if you do that at the moment. I think that was the foundation of essentially the feedback database that we have developed over the years. Then it grew from one column to three columns, then to 10 columns with different priorities. You know how it goes. Every system grows with complexity, especially if it works. Then I think that was used to every week we had a review of all of the team’s comments and feedback. It was quite a fast thing. We didn’t spend too much time. We just went through and talked through what are the assumptions there, and what are the that we’re seeing, and tried to collect what is the signal. We treated this as noise, and we tried to understand what signal we were getting there and if there was a signal. I think this developed over the years. We used a bunch of different tools. We went away from it after almost two years, almost, of using Google Sheets. We switched away from it because they became just too slow.

There were too many records and too many rows and columns in that Google Sheet. I think right now, I haven’t been involved in operations for almost a year, but I think the last time I saw the setup, it was a pretty sophisticated setup with multiple Notion Sheets integrated with information about customers, with information about prospects, with information about companies and people, feedback, and then linking that to a more higher level problem set that we’re seeing. From a process perspective, that became a real process in the company with the sales team involved in working, populating with the customer success team involved in that. There are a couple of integrations that pull data. Then we have, obviously, a dedicated design team and the research team. And, specifically, the research team does a lot of great work in figuring out what the signal is and then doing additional deep-dive interviews with our champions. To understand, Okay, is this the right signal? And do we understand it correctly? And all of those anecdotes go into actually as input for the team that built something. I think that came from Google It was a real sheet in the early days, that became a real process with multiple systems and multiple teams working together, which I think is good because now everyone has full transparency into what are the top problems that we as a company see in the market from our customers and how do we prioritize them? Because it’s also linked to what we want to build, the roadmap, all of that stuff.

Yes, it goes a little bit back into the marketing funnel. I think a lot of our marketing is community-driven. A lot of the marketing is content-driven. I think that’s like, we do sense the trends and we do try to talk about them and understand the pain points and almost what’s the zeitgeist of today’s world. What is troubling the legal community? And one of the things I think a good example is legal Ops, which is today, it’s not such a huge thing, but I think there was a time when it was top of mind for a lot of people. And I think we’ve seen how that overlaps with what we had as a solution. And I think that helped to promote and talk about not just the features of the product, but the solutions that we have for people’s pain points.

I think this goes back to the foundation from where we started. We treat contracts as a thing that lives not just in legal or not just with a lawyer, but a thing that lives everywhere inside a company. That means that whenever we talk about it or whenever we design something, we think about every person who interacts with the system. That means that it’s almost like a salesperson is in the same way a first-order citizen as a lawyer in our platform. They might have a different set of tools that they have access to. They might have a different level of visibility. But the experience that we’re trying to craft is pretty much the same level of goodness. That means that they’re not second-order citizens, so they would not have a worse experience than, for example, the lawyers who craft the contract. That means that when a system is rolled out, for example, to a big team, and there’s a sales team in that team, that sales team, we try to minimize, for example, the number of clicks that the sales team needs to have to do something. I think that definitely from a product standpoint, helps.

That lens of looking at everything in the same way, providing the same experience, whether it’s a counterparty, whether it’s the user in the company, or something else. I think that’s honestly, that’s less, especially for enterprise deployments, for enterprise cases, that is a good product baseline, but it’s just a baseline. I think the magic happens with the amazing sales team and with the amazing customer success team that we have. That is truly the approach that shines there. I think it’s just being decent human beings. That goes to both the hiring practices as well as how you incentivize the team. With that, generally, from what I see, we tend to have generally positive, adequate human beings. They know what they’re doing, so they have a high degree of specialization so they can speak the same language as the customer does, and they like helping people out. I think that means that when there’s a tricky deployment or there’s a tricky situation, the team can solve that. Whereas maybe by just pure software, that could be hard.

There are so many. And with a lot of them, it’s what we went through with some of the companies, we went through actually growing up, almost growing up for years and years, they were a customer, and we have sold their growing pains. I think one of the close ones that I remember is Funnel. Geographically, it’s close to me. I was based in Latvia in Riga, and they’re a Swedish company, so they’re pretty close. I think we had such an amazing champion there, someone who had the right focus of the right mindset and focus, the mindset being an early adopter mindset, not being afraid of using and figuring out how to use tools, and the focus being on really value-delivered. I think with that happens, that’s some magic. When I think about that and I think about the customer, I think we have some numbers on the website because they are a use case that we talk about. We’ve achieved some remarkable steps. We decreased, for example, the number of steps that were required to get approval. They’re an enterprise customer. Their customers are enterprises, so they have a sophisticated contract that they need to agree to.

That contract requires sign-up from the operating officer, finance, and all of that. Usually, it was around 15 steps. Sorry, 50 steps to go from essentially a draft to something signed. We closed the gap to three. I think the more important thing is that the legal team was hated before we solved that problem. Usually, that’s one of the problems that I’ve heard quite a lot. Anecdotally, the legal team is one of the most hated teams in the SaaS business because they are seen as someone who only thinks about problems, and is very slow to respond. After some time, you can hear the same complaints. I think we helped to minimize that friction. The team did a couple of really simple things, integrating into Slack and having those one-click approvals for simple stuff. And all of that just minimizing the number of steps that are required, taking off the stupid work and the bullshit work that usually the Google team was doing, minimizing that, like automating the small things that people just wasted time on, and making sure that the experience, for example, for the sales team is seamless. They always know what’s happening. They always know they have a quick feedback loop as to whether things are moving or things are not moving when things are going to be unblocked.

That feedback loop actually and that communication, actually allowed for the sales team. I think now they quote them as saying that interacting. They love it the most or something like this. Interacting with Juro is a great experience and they cannot live without it. I think that’s an extreme case, but I think it’s a very good one. An example is when all falls into place, so the use case is there, the workflow is there, and the champion is there. We had great relationships, and because of that, the product shined there. That’s my personal favorite, and I have the only warm feelings for the team that implemented that and for a funnel.

Yeah. I think over the years, the situation changed for Juro. One of the things that I think is clear whenever you do a lot of SaaS, you tend to have different companies, and different sizes of companies behave very differently. One of our investors, Christoff Jens, from Point Net Capital has this ongoing theme of creating napkins and then drawing some nice two matrices there. And one of the napkins that he has is just the classification of different sets of customers. And it’s all animals, which helps. And you go from a mouse to an elephant, with the elephant being like 500K contracts a year to the mouse being like $5 per month, something like this. So it’s the distribution. And all of those types and classes of customers behave slightly differently. One of the reasons is that the big ones, the enterprise customers, have a really slow turnaround time for anything. It’s possible to have a situation where an enterprise customer buys, for example, software and then completely doesn’t use it for two years. Then two years later, someone in that big enterprise customer does a review and they’re like, What is this stuff?

No one is using it. We’re going to cancel.

I think that delay cues the numbers a little bit. I think on the flip side, if you go into the smaller ones, I think that the biggest problem for SaaS that deals with small customers is the natural rate of death for those customers for the companies. That means that I think the average was 30% of small businesses are closing every year from all of that are created, which is an insane metric. It’s not surprising that you see that spillover in several users. I think for Juro, we started in the middle, which doesn’t help because it’s not clear-cut. So I think the strategy is not clear-cut. And we started slowly moving up and up the market. I don’t know what the numbers are today, but they were definitely on the 25th percentile, the 20th percentile of the top retentions for a cohort of customers. But then again, I think a lot of customers that we had were smaller customers from 2015 to 2016. I think it’s a tricky thing to dig deeper and segment it.

That’s it. I have two kids. I have a wife, and they’re trying to convince me to get some animals, to get a dog and a cat. I’m still like, I had a dog when I was small, so I know how much work it is. It’s literally like a new member of the family. I’m like, It’s too hard. This was ended up by my wife. She kindly drew me this wonderful dog as a gift. She always reminds me that we should get an animal. That’s the story. I love it. It’s amazing. Go ahead.

Oh, nice.

Yeah, but it is also serious. I think everyone should have animals, especially dogs and cats. They bring our level of stress down. Probably for the performance-oriented of maybe our listeners, that’s important.

There are basically two angles that I have for this. One is purely economical, and the second one is slightly, like the one, almost a social aspect. I think the purely economical one is, overall, I think companies like Juro who are on the slightly boring side of things, not necessarily will hear someone writing a book about it. And those are boring businesses. I am drawn to those. But I think those are the businesses that help people, or at least I would think that’s the mission, help people agree more. And that means creating more economic value. If you can speed up the rate at which you agree, at which businesses agree, I think that means that it’s easier to do business. It’s also easier to do business across geographies, which means that there’s a little bit less risk if you can ensure that the agreement that we have, me and you, ensures that there is no risk there, or at least it’s covered, then it opens up a way for us to trust each other more. I think that’s the economic part is there. I think the social part is more on the last piece the trust piece, where it just makes businesses not worry too much, or at least outsource the worry about the legality, outsource the worry about someone cheating or inserting quotes or breaking the law, outsourcing that via a system that could handle that automatically.

I think a system, the good part about it is that a system is, again, if it has a really good memory, instead of a person, like a person does forget stuff a lot. A system will remember and will help us agree and trust each other more because it will be more transparent. We’ll not rely on a lawyer, for example, who thrives on information asymmetry. Instead of me trusting a lawyer and then you trust a lawyer and those lawyers agreeing, we can agree directly. With fewer intermediaries, more commerce, and more trust, I think it’s a good thing.

Thank you so much. Thank you, Harshit Thank you. Thank you very much.