Building Agile Legal Operations At Scale With Brian Powers

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. Great. Hi, Brian. Thank you so much for, for joining us today. We really appreciate you coming on the podcast. So before I flip things over to Pepe, who's gonna give us a, a better idea of what we're chatting about, we'd love to learn a little bit more about your background and also maybe just a quick recap on, on Markups.ai. Yeah, absolutely. Well, thanks so much for having me. Uh, quickly on me, I started my career in investment banking at Moelis in New York, and then started a company called TemperPack that makes sustainable materials used to replace plastics in the cold chain, and grew that to more than 100 million of sales, over 500 employees. And throughout that experience, I think two things really stuck out. Uh, one, uh, contracts could be quite the bottleneck in growth, and it was always a pain point.

1:02And then secondly, that process was incredibly important to making things more efficient, whether it's on the, the manufacturing floor or in how you deal with contracts. And so after TemperPack, I wanted to help solve the, the contracting bottleneck, and so started Markups AI in 2023 with my co-founder, Sam Shelley, who also, uh, had a career as an entrepreneur, starting a company called Vault Health after starting his career in AI more than a decade ago at, uh, DataMiner. And our, our vision for Markups AI was to create an AI colleague called Agent Marco that could help negotiate contracts wherever they are, so whether they're in your email or your CLM or CMS, uh, and that it would be hyper-customized to you. And so we have a, a team of legal engineers who are former practicing lawyers, who train Agent Marco on all the preferences, whether that's documented or currently just in your team's heads, so that it knows exactly how you like to negotiate contracts and can really intuit what you're looking for. So yeah, we started that in 2023 and been growing really quickly. That's awesome, and that's a very interesting background, Brian.

2:12And well, the whole, the whole topic for this chat will be, of course, CLM, and this AI hype. Like, it's been since 2022, it's been all around the world, and a lot of companies are, like, trying to jump into AI without even know what would be, like, the main reason to do so, right? We were discussing this a couple of weeks ago when we first chatted, that, uh, we've seen this a lot, especially, like, with RFPs, when companies, instead of going to specific use cases, they're just asking, like, "Hey, do you have AI?" Right? But what do you wanna do with that? And one thing that you brought with your intro, Brian, that is, like, very interesting is that one of the main pain points in your previous company were contracts. Yeah. And that's, like, one of the main reasons why you decided to go with, uh, starting a software company focused on contracts, right?

3:09And, and well, like you said, you already built and scaled companies b- before, but what specifically made contracts such a painful problem worth solving and, you know, like focusing now your career on? Yeah. You know, they, they're difficult because they do matter. They guide all sorts of behavior, uh, inside your company and with your vendors and your clients. It's incredibly important. It's where you balance your interests with those of your, uh, partners externally. And so it's not something that you can really shortchange without it coming to, to bite you, and documenting it is just very difficult, so it, it's resisted automation for hundreds of years at this point. I think what makes it so attractive now is that it, it really does feel like a solvable problem. Nothing really new needs to be invented anymore in order to make negotiation 10 times faster with a lot less, uh, pain back and forth. Uh, it just needs to be applied.

4:09And I think most companies have a lot of contracts, so it's this massive problem that's all of a sudden solvable, and that's what makes it such an attractive target. No, I think, like you said, right, it seems like now it's something that is actually solvable, right? I, you know, I think over the past couple of years, we've really seen these challenges that organizations face and technology, uh, being able to support them, going from something that was on a wish list to something that's actually tangible. And we mentioned something when you were talking a- about the application itself, right? We said that "AI colleague," right? And what does this actually look like in the day-to-day for a- an organization or a legal department that's using Markups.ai? Yeah. So I think the, the difference between a colleague and, you know, a general resource is just a colleague that is gonna know your preferences really well.

5:05They're gonna have the experience of working with you, and you don't have to re-explain everything you want every time. You don't have to fear being misunderstood when you give higher level instruction. Y- That should trigger a lot of actions that are very specific to what you want. And then I, I think the other implication is that it works where you work. You know? The companies that call themselves a, a colleague, but then there's a platform that you have to log into and you can only work with your supposed colleague in their environment, that doesn't feel like a colleague. That feels like software from 2010. So I think what we try to do is create a colleague that can work inside of where our clients already work, and that it will know them really, really well. It will know all sorts of preferences that, uh, they don't have to tell it every time. And, uh, I think that's what makes AI really useful and really usable, when it works where they are and it knows what they want. And let's try to think this in the other way around, Brian, right?

6:04Like, you already found, like, uh, you know, like a sweet spot- And I really like this with AI colleague because especially, like, contract managers, right? When they're negotiating the process... Well, well, maybe not contract manager, but actually the business departments, actually, be- because those are the ones who are actually doing the negotiation process. And a lot of the time, you know, like, they knock the door of the legal team all the time because they don't understand legalese and how contracts are actually, uh, still being written. And since three years ago or four years ago, there's a lot of new AI companies or where they, they say they're AI, maybe they're just, like, wrapped with some LLM. But where have you seen those new companies falling short, right? Like, maybe not specific to contracts, but, like, a little bit in general in contracts, of course, right? Yeah. I think the biggest thing is they say AI is easy, and they just say, "It's just gonna work. It's gonna make your life so much better. Just ask it something, and it'll, it'll know what to do, and it'll be amazing."

7:12And I think that just falls so short of what people experience on a daily basis. Some things it can do amazingly well, but some things it really struggles with. And if it's not gonna be consistently good, then it's hard to rely on. So it's not fundamentally changing people's work. It's not taking a ton off their plates off the bat. And so they're just setting this up for s- uh, disappointment, you know, after they sign up when you say it's, it's not actually easy. And, and then they couple that with what's essentially a, a do-it-yourself interface. "Here's a text box. Type in what you want the AI to do, and good luck." We call that the, the good luck approach. And it's just very frustrating. What we see is that AI, uh, if you very carefully train it on, uh, a workflow for a client so that it knows exactly how they like to do things and that it can use the same tools they use, whether it's email or Word or their CLM, then, uh, you can actually get some amazing results, and you can start to depend on it. But it's not easy, and it's way too much work for clients to do on their own. They're already so busy with contracts. If they weren't, they wouldn't be looking into AI.

8:20So adding all of the workflow to automate their work, it's just not something they have the time to do on their own. So we try to be real with clients. We say, "There's gonna be a lot of work, but we have a team of legal engineers that are going to take on the bulk of the work." And so that they're gonna customize it for you using your data, asking really efficient questions so that do the, the difficult work to customize it, so that after that it can be easy. But to say it's easy upfront and clients can train it themselves in the spare time they have is just not realistic. Maybe in 10 years, but we're nowhere close to that right now. Right. And let's think about this example that I was talking about, right? Especially with RFPs, right? When companies are just- Yeah... like looking around and maybe in the discovery stage on what's out there, especially in the AI market. And of course, CLM's been here, uh, it's been in the market since really long time ago. And a lot of companies already have a CLM implemented, right?

9:18Sometimes it was very easy, sometimes it's very complex depending on the company and, of course, depending on how much importance you put into the implementation process. So- Yeah... in a lot of the things, and what happened with some customers is, like, they don't wanna do that anymore. Like, because a, a lot of the times implementation can be a little bit, well, or painful, right? Because there's a lot of details. So I'm pretty sure that you already have some customers that have some CLM implemented in their companies, but how should they approach this? Like, how would you recommend them to look in the market on how to build the RFP so that AI can complement those workflows that they already built with a CLM? Yeah. Most of our clients have CLMs. Most of our clients are, are large businesses and took the plunge, uh, into a CLM somewhere along the last 10 years. And we view CLMs as spaces for collaboration across a company.

10:21It's where colleagues from different departments and different functions can collaborate on contracts in a way that is well documented and clear on the status. And it is the best place to collect documents and intelligence. And so our... when we built Agent Marco, from the beginning, we knew it was gonna have to be able to use a CLM just like our clients would use it. And so our clients just invite Agent Marco to the CLM as a user. Uh, and then Agent Marco can be set up as an approver so that it can get notifications just like their team does. It can download documents, revise them, upload new versions, and then pass it along to the next reviewer. So it really is easy. You don't need an API integration. You can just set it up like you would a new colleague. Uh, and Agent Marco can actually use the browser and do any routine function inside of a CLM that we train it to do. And so that's the new way integration is being done with agents across the board.

11:24I think we're on the frontier in the legal world, and our clients are loving it. It's so much easier to have this be not only in their CLM, but in Salesforce and their productivity tools. Uh, it can have different systems speak to each other without all this different architecture being involved. And so you compare that to all of these AI tools that, uh, are supposed to sit alongside a CLM, but they're highly redundant. They have a lot of the same features as a CLM. They're AI Word plugins, and the idea is that every time you wanna use AI to revise something, you have to download it from your CLM manually, open it up in Word, prompt the AI to do everything you want it to do, clean up its mistakes, and then manually re-upload it to the CLM. And then, you know, now you have your contract intelligence both in your CLM and in this new AI platform. It ends up being really redundant and confusing, and it takes people out of the workflows. What we say is, if you're gonna add something to your system, it needs to use your system. It can't add another system.

12:22And that's really why Agent Marco has fit so well into the existing workflows of clients that already have CLMs and other types of software tools. No, I think that's, that's huge, right? Like you said, I mean, everything is heading towards this simplicity i- is, is really the, the overarching theme when it comes to, I think, how folks are viewing really everything these days. And I think you touched on a good point, Brian, when you said that it was great to hear that, that you're being realistic when, uh, onboarding or working with folks that are still looking at these tools, right? And being realistic and saying, "Hey, you know, there is gonna be some heavy lifting that needs to be done from your team's side," right? From a customer's side, you know, AI is not this magical bullet that, that solves all, all problems. I mean, it's refreshing to hear, you know, somebody else have the, a similar take as the way that we do and right, and how we talk with folks to hopefully keep people grounded and again, realistic, right, when we're talking about these things. But, you know, I'd love to get a little bit of a better understanding. I think we just touched on, you know, how the application works, can work with CLMs.

13:21What would you say that some of the most measurable outcomes are when customers are using the application? What are some of the maybe the quick wins and the longer term wins that they're seeing once this is up and running? Yeah. So the, the most important thing and maybe the most elusive thing with AI technology, uh, for contracts, as pervasive as it is, is usage. Are people using it? And I think what happens with a lot of AI tools is they sign up, and people are surprised that the, the lawyers and the legal teams don't actually use all the new AI technology that's out there. And people are smart, they're not gonna use it if it doesn't actually help. They don't really care if there's some, uh, top-down mandate or people are telling them they're supposed to use it. They'll try it, and if it doesn't work for them, they'll, they'll not use it. They're, they won't continue to use it. So we look at usage, and what we see is that we get to one hundred percent or very close to one hundred percent of contracts that Agent Marco supports. So for example, if we set up a DPA workflow where Agent Marco fully negotiates DPAs end-to-end, what percentage of DPAs come through Agent Marco? And we get to virtual one hundred percent.

14:25And a lot of times that's because we set it up from the beginning that Agent Marco gets every DPA as it comes in, so no one even has to send it to Agent Marco. But we also just see that people opt into it. They, they like sending it to Agent Marco and getting its feedback because it knows their preferences really well. And when they see something that Agent Marco does that they would want to be done differently, they don't all of a sudden have to take on a several hour project of, "How do I fix this?" They just report it to our legal engineers, and they fix the problem once and for all so that they don't have to do the same, uh, checking over and over again. And that's what just makes it so easy to use Agent Marco, drives that usage. People tell us anecdotally it saves around ninety percent of their time in a review once we get everything dialed in. Uh, they're not mechanically making the same changes. They're just reviewing it. So it really turns them from a doer into a checker, and they're just answering things that need to be escalated to them. And then we've had a number of circumstances where clients say, you know, "I was gonna hire another, uh, person on our team for contracts, but I don't need that anymore.

15:27Agent Marco is gonna handle that, so we're gonna hire a different role to help me with something else that I really need." So there's a lot of measurable ROI out there, uh, for when you're actually saving time, and I think everybody is willing to try AI. I think in the next year or two, there's gonna be a lot of churn of AI products that people tried but couldn't show that ROI. And so we're preparing for that reckoning now. Yeah. Yeah. No, I think you're right, right. I think user adoption is, is, is absolutely a, a, a key to, uh, the longevity of, of any type of solution or product, right, working for an organization. So I think that's a major key along with, you know, change management and maybe taking a look at processes, right, of course, before looking at software or making sure that your processes are as, are, are fine-tuned and ready to go before just trying to, to stick an application in the middle and hoping that it solves all of your problems, right? Which, which I think we all know un- unfortunately, that doesn't just work like that. But Brian, this has been a, a fantastic conversation. We really appreciate you, you know, joining us today.

16:30If folks wanna connect with you or learn more about Markups AI, where's the best place to do that? Yeah. Posts on LinkedIn. Apologies to anyone who doesn't wanna see hot takes on legal AI, but you can find me at Brian Powers, B-R-I-A-N P-O-W-E-R-S, Markups AI on LinkedIn. Please connect. And then on our website, which is just markups.ai. You can see more information about our Agent Marco. Uh, you can request a demo. And yeah, we serve all sorts of different clients, Fortune one hundreds, all the way down to small startups. If contracts are a pain point for you, uh, you're negotiating routine contracts over and over again, then Agent Marco probably can save you quite a bit of time. Awesome, Brian. Well, thank you so much for joining us, and thanks everybody for listening to another episode of Contract Heroes. Thank you. 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.

17:29Whether 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 kohoconsulting.com.