Contract AI Secret Tool Lawyers Can't Ignore

Speaker A: Hello, George. Welcome to AI for Business. Um, how are you doing today?

Speaker B: I'm doing fantastic. Thanks so much for having me, Sarah. It's great to be here.

Speaker A: Of course. Uh, today we're speaking to, um, George Haskell, if I'm pronouncing, uh, it right, and he's the founder CEO of Approve this. And it, it says stop chasing approvals. Start getting them. Tell us about what prompted you to start Approved this. What was a big problem that you tried to solve with this.

Speaker B: Yeah, so I've, I've had a, a, A, um, very interesting journey through my career and done a lot of different things in all the different companies. But the biggest motivation to start Approve this is I was managing, uh, operations for a company called mailgun, which sends most of the Internet's email, or much of the Internet's email. And I joined that company at, um, employee 60 or so. And, um, left a few years ago when we'd acquired some businesses were quite large. But through most of my time there, I was approving all of that company's sales contracts. So everything would have to have my signature on it up to, I think a quarter million dollars a year. Uh, and then also every time anyone needed to buy something, uh, they also hit my inbox from, gosh, like $500 to, I think $100,000. Uh, so as you might expect in a business that was growing, uh, very, very quickly, um, my inbox got very full. I, uh, had a difficult time managing things, and oftentimes it was just a rubber stamp for me. It was something I was going to press yes. On immediately. But, uh, that doesn't mean it didn't get stuck in the queue and people would have to wait a long time and they'd be slowed down. So the initial concept for the business was just to have a really simple way to just press yes and keep an audit log and keep the finance people happy and the auditors happy and, uh, my inbox clean. That was the original idea.

Speaker A: Um, maybe let's double click into it. Um, what would you say that Approved this does different from the rest of the tools that are out there? How are they doing it? That's. You figured, okay, if I do it like this, it's going to be faster, better, cheaper, more accessible. Yeah.

Speaker B: So there's. The original idea was really just to make it so that anybody could click on anything. What I mean by that is that really there's, there's, uh, sometimes need approvals from interior customers, people who work for your company. And that's the. My Original use case. Sometimes you need it from exterior people, sometimes you need to take requests from people outside your company. So like imagine you're a marketing agency. You need to send copy for approval and then bounce back and forth like a proofing tool. Uh, so that's one use case. That's one big thing we do differently is allowing external people without needing to create an account or set anything up. And then also um, sometimes you need to accept for like nonprofits you have to accept requests for things from all sorts of places. So really anybody can be anybody. They don't need an account. They can pass information back and forth really easily. And that, that's the big difference that's been successful for us, um, especially in the legal reality realm, which is a surprise for us. About half of our business today, uh, a few years in is actually lawyers passing documents back and forth. Um, which is why we're sort of double clicking into that today.

Speaker A: I actually wanted to ask who um, is your biggest, you know, category of uh, clients. And you answer that question, which is mostly lawyers and probably legal firms. Um, it's very interesting because it does actually I remember that it's always been a lot of effort, uh, awaiting a long time for people to get through this approval process. And when I was a project manager I always thought how I wish that this could be just quicker and faster. Like I wish that I could do that with vendors easier. I wish that vendors had a way to just not have to go through all of this, that kind of manual bureaucracy before they have to go and approve something, the people inside the company to have to go and approve things for them. I'm looking at for people like myself who don't know about how emails work. This is about your previous company. I'm just very curious is it says this company has it helped hundreds of thousands of companies like Lyft, Microsoft, Etsy, Whatnot to send emails daily. What does that actually mean? Isn't, isn't like Microsoft. I know it has out Outlook for example. How is that different? And then how does that connect maybe to your technology? If it's kind of, if it's relevant to that.

Speaker B: Yeah, yeah. We'll go back into my ancient history for, for a second including uh, before, before I joined. So mailgun is a, is an API driven company. So it's how developers send send emails, either transactional emails or um, transactional emails or other types of emails. And so really competitors in the space are like sendgrid or mailchimp to a certain extent. Maybe people might be more familiar with that one. But really when you're sending hundreds of millions of emails, potentially a day, or you know, millions of emails in an hour, like maybe a company like Starbucks would do, you want to make sure that they always get to people's inbox, that they show up reliably, that, that they, they get to where they're supposed to get um, on time. And so really one of their first killer features was inbound emailing. And that's one of the features I always really liked. It's actually one of the things approved this does really well. So I mentioned at the beginning that my inbox was always full of approval requests. So one of the things that we do with approved this is we take inbound email boxes, inbound emails, and parse that message and that document into an approval request that has structure and auditability for people to manage their approvals. Which um, is also how we sort of feed into the way we integrate AI into our platform these days with making sure that all the information that you need is there and also keeping track of whether this is something that you're likely to approve or whether you don't even need to look at or whether you just need to take a brief look at it or dig in more deeply.

Speaker A: Got it. Uh, um, so besides the um, beside legal um, entities who this product really is for today and they, who can actually benefit from using your product, you'd say.

Speaker B: So there's um, besides legal entities is a tricky thing to say because that's really where, where we're pushing into and I'll talk about that in a second. But our biggest non legal customers are really nonprofits. And then um, we have a third category of people that are um, for lack of a better term, and I hope, I hope they don't hear me call them this, but they're, they're Luddites, they're Luddite organizations. And so I'll give you an example. Um, one of them is a forestry company. And so I'm the first tool they're using that is not pencil and paper. And that's actually like one of the most common things that we get from companies uh, that, that are currently today filling out information on a piece of paper and putting it in Judy's inbox and waiting for Judy to pass it on to John, all on a piece of paper. Um, those are the types of companies that find comfort in us for a lot of reasons. We're very happy to have those customers, but that's not um, necessarily the type of customers I always want to have and a space I want to grow into. Um, so now we're sort of double clicking into the legal aspect where because we have these legal customers anyway we're dealing more and more with uh, instead of just passing back and forth like leads or um, contracts we're actually getting into the contract editing space just a little bit. Not a full redlining tool but just the capability to look at a contract and say hey, which parts of this contract are going to need to be approved by which different people in my organization and routing off to the appropriate spaces. So good for internal general counsel, uh, or sales, uh, teams especially to interact with legals.

Speaker A: Right. And that is I'm assuming the space. Well you actually touched on that this where AI comes in and helps kind of with reasoning and understanding who does it belong to or what role it belongs to every section of the contract. Right?

Speaker B: Yeah, yeah. There's, there's two big aspects. The first one is definitely in the decision making. What's usually done with this, how does this request compare to all the rest of them? So some of that is logic trees and some of it is like, is sort of evaluative. Um, and that's where we use a couple different large language models depending on the purpose. And then the other item is the routing, uh, looking at a clause in a contract that's changed from A to B. And does that need to go to the finance team? Does that need to go to the legal team? Does that need to go to an operations team or you know, uh, development team figuring out who needs to, who needs to look at this and say it's okay.

Speaker A: Right. And does it also um, kind of follow up or kind of send okay notifications? Hey, like Jack in the legal, in the finance department they looked at it and then you know, it goes back on the workflow. I'm assuming that kind of there is a, there's a way to understand that this workflow is actually kind of automated in a way. So the follow ups are easy. I'm just kind of maybe giving you a feature idea here.

Speaker B: But yeah, no, no, absolutely. So yeah, we like to say it's a way to not have to hound people. You let the system hound people. You can set it up to bug people every day, every couple days, basically bother them until they press yes or no. Um, and so yeah, that's absolutely built into the system. The notifications are all super customizable really can sort of. We started out as this very simple sort of thing and one of the, one of the problems as an Entrepreneur is that the more you talk to customers, the more customers ask for. My tool's very horizontal, right? This is historically we're very horizontal. Everything we're trying to be, uh, as survey monkeys is to surveys approve, this is to approvals, little slice of approvals for everybody. When you build that way, it's just a never ending roadmap of things that come in and say hey, I needed to do this and this and this and this. And some of those things, especially now with, with some of the coding tools available, they're so fast to build. You end up just getting the huge, A uh, huge roadmap of stuff that's shippable, that you can get out the door and it works. But uh, what that means is that you stop being, you become a very complex horizontal tool. And the more complex you are, the less useful you are as a thin tool. You stop being a thin tool that's good for everybody and start being a, you know, a vertical tool that um, most useful for very specific people. Which again is why we're focusing on that legal aspect. We got, got the people there already, they like it and um, so that's why we're double kicking on that.

Speaker A: So question. How about for like companies like other businesses like myself that we want to sign contracts. So I'm sent a contract, I'm looking at it and then I will have maybe my lawyer looked at, look at it if it needs, if it's complex enough for a lawyer to come in. And then after that, um, for kind of smaller simpler businesses, they um, want to have the back and forth. Maybe the lawyer, my lawyer and they lawyer, they want to have that conversation or I want to kind of be looking into some pieces and say hey look, we want to commit to that. This is something we can, this is something we can't. So how did, is it solved also for the external, um, kind of piece as well? I know that you touched on it that it's also for external but like you know there's these redlining. Um. Is it called redlining when they. Yeah, yeah, yeah. So that's worth, so to speak.

Speaker B: That's actually the core use case. The, the, the um. And just to be clear, so we sort of wandered around a little bit around, around my roadmap and product. This is um, we still have all the business process management, the normal approval stuff. This redlining stuff is actually coming out this coming week. So we've got a couple people testing it out. But it's, it's sort of brand new. So um, here's how it works. Uh, basically the redlining is the core use case. You receive a contract, you want to bounce back and forth with an external, uh, entity, whether someone you're selling something to or someone's selling something to you, and sort of figure out whether or not it's something you can accept or what are the things that are different that you would normally accept that you're not going to accept this time. What are those pieces where there are gaps? Um, and then who needs to approve them? So it'll go in and it'll look at every clause. And if sometimes, if it's you selling something, it'll compare it to what your standard contract is. So what are the things that I always agree to? What are the things that need to be looked at? And it'll look at the decisions you've made in the past and say, okay, this is something. Like they want to change the jurisdiction from California to Delaware. How boring. Like, this is something I don't care about, or it's something I care about a ton and I always want to do that. Or I don't want California or Texas. I want it to be somewhere else. Um, so it gives you the capability to quickly pull those items out and really supercharge the billable hour that you're paying to your attorney, especially if you don't have, like, someone who's working just for your companies. A lot of us don't. Right. So it supercharges the billable hour and makes it so you only send the stuff to the lawyer that, that the lawyer actually wants to look at. So really, it's an interface. That's the idea to be an interface between the legal aspect of your business, especially if it's an external, uh, and your internal team. And then, of course, the next step is figuring out who needs to approve it inside finance or operations or, um, anyone else.

Speaker A: And does it also kind of. Is it like, kind of like a workspace that gives visibility to all three or four of us that we can see, like, what's happening, where the document is at now, for example, like, if it's waiting for my lawyer, you know, what part is he looking at things like that again? Maybe I'm just like, like throwing ideas around. But it's. It would be, it would be nice to have those kind of, um, you

Speaker B: know, you've got it exactly right. So. So really, um, the exact statistic, I believe, is 76%. It's 76% of lawyers still do most of their work inside word. Um, and so to get to get technical because a lot of my background is as an engineer as well. Um, we sort of reverse engineered the way that Word stores documents. It's an open source. Docx is an open source format. It's like a zip file or other documents. So we, we looked at that format and we said, all right, well we, it needs to work like Word because lawyers use Word today. And so the first thing we decided is this has to be interchangeable with Word so you can drop contracts in and out of Word and all the red lines that you're working on can just. If your lawyer doesn't use our tool, that's perfectly fine. They can use Word because that's what a lot of lawyers still use. Most lawyers still use. Um, but also inside the document you get everything you'd see in the redlining view of Word with all the crossouts with the current state of the document, along with all the comments, which is again interchangeable with Word comments, just like you have in Word or Google, Google Documents or anything else. Uh, along with our, uh, internal. Here's the section that needs to be approved. Here's who it's waiting on. You know, here's the status of the document, which, you know, if you think about, especially in sales processes, something can really have an roi. Oftentimes when you get to the contracting phase of a deal, things move very quickly until you get to redlining, because that's the moment when salespeople lose control. Like up in the, up until then, they're setting the next steps. They're going through the process, they're filling out salesforce. You send it to their lawyer and it just disappears. They never see it. It sits in the lawyer's queue until they get around to it. Right now those become very tactical, actionable steps. Before I was waiting for the lawyer to review, waiting for the lawyer to check with the CFO on about some term. Now it's, here's who needs to look at it next in a way that doesn't take any control away from, from um, the legal team, but, uh, does provide visibility to everyone else involved.

Speaker A: This part of it, it's as a, as, as the business, the business owner. It's very relieving and nice to have because I want to see what's going on. I want to see where is it? Is it stuck? Can I help everybody? Anyone? Can I like push things? Can I like, you know, just the visibility and transparency. It helps a lot, I think, from my perspective. Um, and for the lawyers. So everything is being more seamless. It Seems, um, it kind of helps with the speed of things, of process. The process is being more streamlined. So it's basically, um, saves time for people to be able to kind of work, um, with the documents. And then do you have in mind maybe that even for the reasoning part of it? Okay. Like if we are replacing Delaware with California or vice versa, and that's important for us. In the example that you mentioned, like, maybe I'm like, uh. Because I see that in my contracts, I'm like, okay, this is go. This, uh, client would like to kind of use this jurisdiction for, um, their, you know, legal affairs. How is that different from mine, for example? Um, again, as a business owner, Because I think lawyers know that already or.

Speaker B: Yeah. Why does it matter? First of all, um. Well, second of all, third of all, I'll say very, uh, importantly, I am not a lawyer. And, uh, we. This is a very tricky space, right, because there's very strict regulations that vary a lot based on your jurisdiction about, uh, what legal services you can provide, what you can say as a software platform, what you can only let a lawyer say. Um, but the idea is to make sure that we're providing enough context for business owners to make decisions and know when they need to reach out. And so that's. That's a dial we're turning back and forth because I think it is fair to say, well, here are the things that. Here's why companies do change from one state to another. Here are the things you are going to want to think about without providing, you know, a recommendation, uh, just to provide that little extra context, which is really interesting because what I've heard a lot of people say when, when we talked about this and got them to try it out is, hey, I'm just throwing it in Claude today, or check, I'm just going to throw my contract in there and see what matters. And that's really where the structure of the business that we've already built, uh, the rules, the. The deterministic parts. You know, this isn't just. This isn't just a contract that you're going to throw in and get some general advice back. This is going to turn the contract into actionable things to look at. That's where the rest of the application we build becomes really, really helpful.

Speaker A: But at this point of time that you're about to release new features coming up, um, do you want to talk about those features or do you want to still keep it a surprise?

Speaker B: It really is that redlining stuff. The redlining stuff goes live next, uh, week, uh, We've been testing it and building it for, gosh, longer than I've ever built anything else. About two years since we built sort of the MVP of this and thought about what it could look like and bouncing it back and forth with lawyers in my network and companies in my network and sales teams especially to, uh, make sure that it's worth, worthwhile. Um, but that's the big stuff that's coming that we're really excited about.

Speaker A: Well, congratulations. It's going to be the second week of May. So this is the first week of May right now, and your release is going to be in the second week of May for, um, listeners who are interested in this to try it out.

Speaker B: Yeah, it's, it's, um, going to be right@redline.applove this.com. i believe in keeping names very simple. I, uh, love domain names that say what. What you do.

Speaker A: Yeah, exactly. They're like, right to the point.

Speaker B: Yeah.

Speaker A: If you want to say, um, like, what is, what is a big challenge that you have right now that, um, keeps you awake at night, so to speak, or just, you know, occupies your mind around about the product or you're building, or maybe it's, um, about what the ICP needs. What is your biggest or a big challenge that you have right now?

Speaker B: You know, um, I have a background in engineering, as I mentioned, and operations, um, but I spent several years as a, as a product manager as well. And so, you know, balancing what can get built with what should get built has gotten so much more difficult with, um, AI coding tools as they exist today. It's so much with a small team, it's so much easier to build anything that anyone ever asks for. And so being really laser focused on who your ICP is, which is really what we're focusing on doing right now, um, has gotten much more important than it was. Similarly, the way that, um, you know, what people call the Sasspocalypse, which I don't really believe in, but the idea that all software will die because everyone can just vibe code their own tools, uh, I don't think that will happen. But I do think, um, it makes it more important to go vertical like we're doing as well. So the two answers would be that really deep focus on ICP and making sure that we've got that right, because the more specific you get, the harder it can be when you're wrong. And then, um, making sure that you're building the right features for that icp.

Speaker A: Right. And to your point about Vibe coding, I feel, yes, there can Be a lot of what I'm seeing. There's a lot of products that can be built, but which one's built? Right. You know, there's a lot of. There's a lot of nuances that goes into building your product, which it's not just about how it looks like. It's not just about, okay, it works, but is it secure? Is everything. All the nine yards of software engineering, which I'm not a software engineer, is it considered in building that product? I think we still do need software engineers to kind of, um, at least have an oversight of what's happening, um, when we're building all these new features and these new products that are coming out. Right.

Speaker B: My favorite story about that is many, many years ago, I worked for the Walt Disney Company, and, um, I was just a retail manager, but there was a problem that I found in the way that we restocked our shelves. And so I wrote a piece of software in Visual Basic, ran inside of Excel, and it helps our floor stockers figure out what needed to go on the floor. And it turned out was really successful. It saved the company millions of dollars. And I, you know, I got a night job of, like, maintaining that piece of software and then helping it go into other parts of the organization to the point where it was running hundreds of millions of dollars a year and saving an uncalculable amount. Um, I changed jobs and it disappeared in a week because I wasn't there to maintain it anymore. Um, and I think a lot of the Vibe coded stuff sort of fits in the same bucket. Like the. What software is isn't just the code or how it works. It's the maintenance, it's the support, it's the releases, it's the curation of what it should do, what it shouldn't do. All those things are so important in business, and they're not captured by the vibe code stuff.

Speaker A: That's correct. So your focus on your ICP right now is lawyers and legal firms, for the most part. If you want to look at your product, let's say, because we can't say five years now in the age of AI, uh, if you want to say, if you look at it in, like, I don't know, a shorter period of time, let's say two, three years from now, um, is that still going to be your only focus, or you think that you want to expand it to other technologies, types of businesses that can use your product or benefit from using your product?

Speaker B: I would love to continue down this path because I've spent so much of my career working with both sales teams and legal teams. So, um, I think that, uh, this new set of features is, uh, very tailored and puts us on the path to becoming what's called a contract lifecycle management system, uh, competing with bigger systems like Conga or Ironclad, which are, you know, huge enterprise things that manage contracts and obligations and renewals for, for companies. Um, and so when I look two years down the road, that's, that's really where, you know, assuming that this catches the way we'd like it to catch, and I absolutely think it will, that's the product that, that we'd like to build. But also as a small business and as, as a, you know, hungry entrepreneur, it's my job to figure out if it's, if that catches, you know, and, and, and it works the way we think it will, then we go full, as fast as we can, as many resources as we can into making going as far down the path as we can get. Um, but if it doesn't, and it's, you know, every indication is that this won't happen. But, but, you know, we keep, we keep finding that next vertical that, that may make the biggest difference for us. I think the biggest thing, though, is the biggest change that AI has forced in our business is making sure we do focus on the vertical, whatever that may be.

Speaker A: Right. And it's always a product market fit question that, like, startups need, um, to kind of keep finding and keep proving which space is the best for them. Either add it or, you know, develop into it. It's, it's a, it's a, it's a journey. Right. I would say an iterative journey.

Speaker B: Yeah, absolutely.

Speaker A: Um, the question that I always ask my kind of, um, one, two, last question is, is there anything that you hoped that I would ask, um, uh, do you know, or anything that you want to talk about your product that I didn't ask a question about, or yourself?

Speaker B: No, you know, um, I think that, um, I think that you've allowed us an opportunity to cover everything we wanted to. I don't think I have anything. No, thank you.

Speaker A: Okay. And how can, um, people find you?

Speaker B: Super easy. Yeah. Um, I'm George Haskell on LinkedIn. There's not that many of us, uh, uh, but I'm the guy with red hair who talks about AI too much. Uh, and then approve, um, this dot com. Approve, uh, this dot com. And then if you want the, if, uh, you're interested in the Redline stuff for legal teams and sales, uh, teams that work with legal teams. That's@redline.appoice this.com. no hyphens, no fancy stuff. Just approve this.com.

Speaker A: okay. Amazing. Well, thanks so much. I'm going to add all of these links. Thank you for giving us time today.

Speaker B: Great. Thank you so much, Sarah.