AI Meets CLM: Kyle Myers on Building the Future of Contract Intelligence

Made with speech-to-text software, so a few words may be off.

0:00You're listening to the Contract Heroes Podcast, your one-stop shop for all things contract management. And now here are your hosts, Marc and Pepe. Hello everyone, and welcome to another episode of Contract Heroes. Today, joining us from IntelAgree is Kyle Meyers. Kyle, welcome to the show. Thank you so much for being here. We're really excited to learn more about IntelAgree and some of the strengths of the solution. You know, I think before we jump into some of the questions, uh, it'd be great to get a little bit more information on your background as well as, uh, IntelAgree's. So yeah, well, the floor is yours. Yeah. Thanks, Marc, and thanks for having me. Big fan of what you, uh, and the team are doing over at Koho Consulting. So I'm Kyle Meyers, uh, chief product officer here at IntelAgree. I've been with the company, uh, since early days. I joined in, uh, early twenty nineteen. The company actually started in twenty eighteen.

0:54Original sort of genesis for this company was to be an AI native end-to-end CLM solution, and that is... The world of AI has evolved rapidly over the past ten years, but especially over the past twenty-four months. Uh, so it's been a, a really wonderful journey with this team here at IntelAgree. Prior to my time at IntelAgree, I spent, uh, a few years at KPMG doing back office transformation, robotic process automation for a lo- a number of large companies, uh, and had a great opportunity to come over here to IntelAgree and work with a great team. That's awesome, Kyle. And can you tell us a little bit on what inspired IntelAgree to focus on AI in the contract space? Because as you well said, like ten years ago it was like more clunky and people started like, you know, working a lot. It was like really complex to start building those models. Then after ChatGPT, like everybody was just like it was going mainstream, right? People were asking everything from creating songs based on specific topic to now creating images, but like what was your inspiration to focus specifically on contracts? Yeah, it's, it's actually a pretty interesting origin story.

2:10Uh, so before starting IntelAgree, our co-founders and a, a few members of our technical leadership were approached, uh, at a previous company by a big four accounting firm who believed AI and ML were gonna disrupt audit practice. So this led them to a, a pretty significant three-year engagement where they introduced machine learning to the audit process. Uh, in the project, they were leveraging a, a tool called Kira to read contracts and financial statements. Part of this project, you know, licensing Kira was over two million dollars a year, forty-three developers helping to build tooling around it. Huge success in building out this platform, still in use today, obviously with some evolution over time, but still being used today. But that kind of cost and, uh, implementation is, is, is really a non-starter for most companies, uh, to think about beyond just the licensing fees, paying forty-three developers a year to build this kind of tooling, this just for one particular company. So realized that cost is, was, was pretty much a non-starter for most companies.

3:11Also got to see that a lot of the amount of time that auditors were spending auditing contracts for RevRec, and then kinda took a look at the front end of the contracting process, and no one was in love with their CLM at the time. Uh, it was still a pretty immature space. So that the co-founders, uh, really had this epiphany that if they had a platform that had some ML built into it that was native to the way that it was working, they could create, negotiate, e-sign, manage, and analyze contracts. They'd have a product that was pretty well received in the market, and this was really validated, uh, shortly thereafter as the company was getting off the ground, as we saw, uh, companies like DocuSign go out and acquire Spring CLM and, and Seal Software, really bringing all three of those things together. So that's kinda how we got started here. Yeah, that's awesome. But I- Yeah, that was like a very good summary of the CLM space in the previous eight years, and that's very interesting.

4:06Yeah, I mean, I was also like as a spectator on how all that CLM space started growing, and DocuSign buying Spree with Spring CLM and all the AI platforms back then. And one thing that I'm particularly curious, Kyle, like you describe IntelAgree as a native AI CLM, and, and you know, like we've been working on the CLM space for some years already, and usually AI, it's still considered like a little bit on the, as an upsell, right? When e-either people are, are, already have some CLM already im-implemented in their org-organization. A lot of the times they just want to jump into the AI train because ev-everybody else is doing that, and they ask us like, "Hey, can we do AI?" And then when we ask them like, "Sure, but exactly what do you wanna do?" And that's like a little bit on when you start getting like a little bit on the frictions, right? Because people want it, but they still don't know what they exactly want.

5:09So I'm curious about this, Kyle, like what does native AI CLM and what does that really mean in practice for your customers? Yeah, so for us it means that the, the user doesn't have to do anything outside of their normal day job to be able to get the benefits out of AI within the tool. So when we started with just machine learning models, wanted to make it as easy as just sort of using your mouse as a digital highlighter as you were marking up contracts as you would, you know, as a lawyer would, would normally review a contract and maybe use a highlighter to say like, "Oh, these are the, the twenty things that I really need to pay attention to in this contract." Being able to do that in the platform, that automa- and, and then store those values so that they could be audited, reported on, used for approval workflows, things like that, those were also creating new data sets that could be used to train new machine learning models. There was no extra effort that was needed in that process. You know, when we originally rolled out our own off-the-shelf models- We started with publicly available contracts, SEC contracts, things that were available in the Quad database, things that there would be no concerns around privacy or, or confidentiality around as we were really trying to drill in on, on common themes for these.

6:25But at the same time, we wanted to build out the ability for our customers to train their own models too. So while we ship a hundred out-of-the-box models, our customers have trained twenty-seven hundred of their own models over time that are either refinements of our models or things that are completely custom to the way that they do business. From industry to industry, there's a lot of deviation, but even from company to company, there's, there's deviations, and there's things that might be particular to the way that they do business and they wanna be able to train on. So legacy machine learning days, we wanted it to be really easy to, to train and add those new data points. No data scientists exist at our, in our customer base, right? These are, these are legal folks, contract admins, a lot of users that are good contract folks, but, but no data science background. So we had to make that easy for them. Now, fast-forward to today, when we have generative AI in the, in the platform, we wanna make sure that we're able to weave it into the way that they're already comfortable using IntelAgree. So whether that's just having a, a pop-up modal that's making it, uh, easy on a page that they're in, meeting the user where they work through the, through our Word add-in, so that while they're redlining a contract, they're, they're able to leverage the goodness of their contract playbooks, get recommendations for redlining.

7:35We want it to have all of the context of all of the contracts that they already have in the system, the way they've negotiated previously. Uh, so for us to be AI native means you're not going out to different tools, you're not going out to different systems, you're not constantly adding context into the question that you're asking or the task that you're trying to ac-accomplish. We have all of that context already, and so it's really important for us to be able to bring that to the user and then extend that to way that, to the where they're working and how they're working. No, and I think, you know, that's something that we see a lot of success with is kind of that, that crowdsource model, right? Allowing customers to feed off of what the team has already built, and as they continue to grow and scale, right? Hopefully leveraging what you have already, uh, created for them. But as they look into the future and they want more things that are more, uh, configurable and customized and more pointed for, for their specific use cases, you know, they have the flexibility to get in there and, and build those things out.

8:29I think that's a fantastic functionality of IntelAgree and, and being able to, to crowdsource those types of things. It's, it's great to see. And, uh, Kyle, you know, the next question here is, is around Sage Assist, right? And how does that kind of represent the team's approach to, to GenAI when it comes to the, the contracting process? Yes. So Sage Assist, uh, i- is a suite of modules that provide sage wisdom and advice to our customers as they're, as they're navigating through the system. Those modules sort of build upon each other. Um, so we have basic, uh, NLP functionality for help, compare, and summarize, so helps with adoption. So if a customer just says, "How do I, how do I create a contract?" Right? They're gonna get references to the tooling and, uh, documentation that already exists in the system. They're also going to be able to build upon the draftable comparison tool that we use in our system to see what's the common language delta on all of these redlines.

9:26Like, what-- in plain speak, if I had to represent what are the three or four things that are different between these two versions, what does that look like? From there, we go to, to chat, and that allows them to leverage contract advice, which means that not only within IntelAgree, but also within integrated systems, users can ask questions about individual contracts and get those answers without having to, to bug legal or, or to go through a process or to go try and track down contracts in a repository somewhere. They're able to get that information very quickly wherever they're sitting at that point. It also helps to, with those complex queries, right? Every... Or not every user claims that they're a reporting expert, but we wanted to make it as easy as possible for them just to type in like, "I need to see all contracts with payment terms greater than thirty days that expire in the next ninety days," something like that. That could be a complex query to put together in a tool and adding different operators in. They can just type that in, and they're gonna get the search results back that they need.

10:21But then from there, we build to negotiation, which is that's really where a lot of people wanna be right now is in augmented redlining, right? Based on some playbooks, based on some rule sets in our platform, based on contract types, how do we make sure that we're redlining this in a way that's consistent with the way that our company wants to do this? Possibly across a large legal organization and, and possibly across different entities and maybe even global geographies, making sure that there's a uniformity around the way that they're doing that. In all of those, it's really important for us that we're showing our work, right? So we don't wanna just give them an answer. We don't just wanna say, "This is the red line you should enter." We wanna say, "This is the red line you should enter, but this is why we're, we're making that recommendation," so that they have comfort. This really helps to build trust. Um, I think trust is a concept in, in AI outputs that's continuing to evolve.

11:14I think Chat-- the ChatGPT, the Groks, and things like that have made it easier for people to have more familiarity with generative AI and maybe a little bit more trust, but hallucinations still exist, right? And so they wanna make sure that, like, what, what was the rationale for why this, this recommendation was made? So it's really important for us that we always show our work to build that comfort. And then we think about data extraction, right? I, I know, um, in a previous podcast y-you had some folks on talking about la- mass migrations into the platform and generative AI and data extraction, right? That's really important for our users when they're coming into the platform. We wanna make sure that we're able to look at complex data sets, you know, maybe twenty, fifty thousand contracts. Maybe they're not all contracts, maybe they're not all fully executed, maybe there's duplicates, right? How do we make sense of those very quickly and get those into a really strong repository? But then as new-- the implementation process can be challenging, es-especially for customers who haven't done this before.

12:14They know what they wanna track to a certain extent, but, like, what if there's 10 or 20 other things they wanna track down the road? We've built tooling to make it so that, okay, you wanna go track one new thing? You wanna go track 10 new things? Give us a name and description. Give us an example of that. We'll go find it in your entire repository, but also set it up so that as new contracts come in, we're tracking that thing, too. It's really big in the M&A world, too, right? So as they're bringing in... They, they acquire a new business, now they've got 5,000 contracts that they didn't sign originally, but they've got the obligation for now, really helps them to, to better understand what's going on there. So those are some big things for us in Sage. And then the one that was released earlier this year is around risk management. So allowing customers to look at their, their risk profile across all of their contracts, whether it be by party, whether it be by contract type, whether it be across their entire profile, but helping them to understand throughout the negotiation process how their risk profile and tolerance is changing based on continued negotiations. And the last thing that I'll add on Sage is really it's important, everybody wants generative AI, and I think every large organization has some sort of a mandate to, to bring AI into the system.

13:21Most of them don't know how to get started, right? They all have the b- the blank paper problem. So with every new tool we release, we put, uh, tooling in place also that helps them to get started. So whether it be looking at their templates, whether it be looking at their old contracts and making recommendations for this is what we think your playbook maybe should be, this is what we think your risk profile should potentially be. It's always easier to edit than create, so we wanna make sure we're putting tooling in place to make that easier for them. Yeah. That's awesome. And I think I was actually thinking about that a lot of the companies are starting to use Gen AI, especially for the red lining part, because back before there wasn't any Gen AI in place for the CLM systems, the negotiation phase was the one that was... That wasn't really easy to automate because you, you can automate the pre-signature phase, all the intake pro- process, approval workflows, e-signatures, maybe tasks and, and obligations when you think about on the post-signature.

14:22But a lot of the times, the, the negotiation process is really hard to reduce because that was, like, more human-centric. It's, uh, depending on the risk you are eager to face or depending on the counterparty, if they are n- like, it's all in or lose, it's... A lot of the time, like, lawyers have that kind of way on doing things, especially when they need to negotiate a contract. But one of the things that I've seen, especially with these new Gen AI tools and when they can track the changes in the clauses during a cer- a cer- cer- certain period, you don't have to go to contracts 10 years ago, maybe just in the past 12 to 24 months. And if you've seen that most of the time, once every three times, you're getting pushbacks in certain clauses, maybe you can have more data on how you can update those templates and then starting in, in something in the middle, right? Because a lot of the times, if you go back to the basics, especially with the templates, that can help you a lot to getting less pushbacks on the negotiation process.

15:31And if you can use AI to analyze, you know, 100 contracts instead of putting five lawyers to doing that, w- wh- they would spend like, I don't know, 100 hours each one just to get that information, I think there's really good value there, Cal. That's really cool feature. And having that as a background, Cal, I wanted to ask you something, like, more into the sales perspective, right? Because usually when you want to sell a CLM platform to another organization, the person that you need to convince is usually, you know, of course, general counsels, but some of the times legal doesn't have enough budget, like procurement. If you want to the procurement director, but the procurement director needs to show some numbers to somebody in finance, either the CFO, depending on the size of the company, or somebody working on finance. Like, what would you say, like, how does Sage Assist help teams to reach value faster, whether it's legal, sales or finance? And, like, how can you sell that?

16:30Like, how can you put some numbers there on how they can reach the ROI? Yeah, absolutely. So you're right. While the legal persona is often the persona targeted by CLMs, they, they are most frequently viewed as a cost center and don't have, you know, a ton of, of sway in the overall budget. Ultimately, getting to the CFO is important to us, but we wanna work with the CRO. We wanna make sure that we have the, the sell-side contracts. We wanna work with the, the chief procurement officer. Uh, we wanna work with risk and compliance. We wanna work with IT, quite frankly, to, to make sure they understand the safety and security around their contracts because they are the, you know, the most important documents to the way that companies do business. You brought up a good point around sort of contract governance as you think about changes to clauses over time. We can make recommendations that, "Okay. Well, you're always giving to 45 days on payment terms. Why not start there?" At the same time, they may say, "Well, it's strategic. It's a part of the negotiation," right? Like, "That's the way that we just wanna give it up."

17:29A- as we think about really demonstrating ROI, for us, it's, one, time to value and user adoption, right? How quickly can we get you up and running on this new tool, which is why we've built a lot of tooling around data extraction. But two, how can we turn those contract cycle times down, right? How can we make sure that we're pulling the right people into the process faster and consistently? How do we make sure that that's an auditable trail? How do we make sure that the delegations of authority that are in place are being adhered to? And so all of those things, right? That's why having a, a full CLM tool that has AI built in versus trying to use a point solution for, like, maybe, like a, like a DocuGPT just for, like, this one editing thing, you have that full visibility into your entire contract portfolio, the entire negotiation life cycle. You have the auditability, you have the trust that goes into that. That's really important for us. Honestly, a, an easy way to demonstrate value for us is oftentimes people forgot that they had a contract that was gonna auto-renew, right?

18:29So now they're on the hook for another, uh, $100,000 on a contract that they forgot they were gonna auto-renew on. Or on the other side, they knew it was th- they knew they needed to negotiate, you know, better business terms on the next turn, and they just forgot. They've got 20,000 contracts sitting in SharePoint somewhere, and they just didn't know that they had it. So making sure that there's really good rules around all of that, making sure that there's a workflow that's gonna be easy to follow and easy to modify over time, but then making sure that they have a really good UI that's going to, to call them to action very quickly, uh, that's going to make the process easy, that's all really important for us as we're demonstrating ROI in, in the overall process. No, that's great, Kyle, and I think that, you know, that's going to resonate with a lot of folks, you know, out there that are starting this process and hopefully listening to this podcast. So I think to wrap up, uh, we, we like to ask folks if they wanna learn more, if, if the folks that are listening wanna learn more about IntelAgree, they wanna connect with you, uh, where's the best place to, to do that?

19:29Yeah, so, uh, certainly IntelAgree.com. There's even a, like a preview on that site, so if somebody wants to drop a contract in and, and just see what Sage Assist would do for them on an individual contract, that's a great way to kinda try before you buy kinda thing. Great way to learn more about our company as well, and you can find me on LinkedIn, and happy to connect with anyone. Regardless of, of whether you're gonna buy IntelAgree or not, I think I'm happy to, to give some best practice advice as to what you should be looking for in a CLM, and I think it's a, a space that's continuing to grow and evolve and continuing to rapidly benefit from the, the advancements in AI right now. Awesome. Well, we really appreciate you, you joining us today, and, uh, we're excited about our, our future partnership. And thanks everybody for listening to another episode of Contract Heroes. Thank you, guys. This episode of the Contract Heroes is sponsored by Koho Consulting. Koho helps organizations of all shapes and sizes find, implement, and manage their chosen CLM tool.

20:26Whether your project is buy or sell side led, Koho can help you navigate the waters of the CLM space to make sure that you're getting a product that's going to fit like a glove years to come. If you'd like to learn more about Koho, visit their website at C-O-H-O consulting dot com.