Served, Then He Bought the Company: What Law Firms Get Wrong About AI

Collin McKee

Welcome back to the Endeavors AI Podcast. I'm your host, Collin McKee, founder of Endeavors AI, where we talk about the intersection of artificial intelligence, law, business, and the new economy being built all around it. Today's guest has one of the more unusual founder stories that I've come across. He's a West Point graduate, Rice University MBA, and a career at Citi. And then he gets served legal papers and ends up buying the company that served him. He's since rebuilt that business from the ground up, layering AI and automation into nearly every part of the back end while keeping the human side — the actual service process — exactly as it's been since the 1800s. He's also actively acquiring other companies in the space right now. Raj, welcome to the show.

Raj Kankaria

Hey Collin. So great to be here. Thanks for having me.

Collin McKee

Awesome, thanks for being here. So let's start where most people wouldn't expect a business story to start. You got served papers and somehow turned that into owning the company that served you. Can you tell us a little bit about that story?

Raj Kankaria

Yeah, I can. I was laid off from my last corporate finance role — the project was a dud, they shut down the office, let go of everyone in Houston. At the same time, I had an unhappy wife, and she filed for divorce. And there's a guy that showed up at my door. Now, I was in a spot where I had to figure out how to make money quick — I was between interviews and needed to figure out something to do. This guy shows up at my door to serve me papers. I asked him how much it cost to serve me, and he told me a hundred and something bucks. And I was like, wow — this is a business. I could do ten of these a day and continue paying my bills until I found my next role. And that's exactly what I did. I went and got a process server's license and served about five to ten papers a day, got all my bills paid. But I fell in love with it. I loved driving around Houston, loved knocking on doors, serving papers, getting that dopamine hit every time someone got served, getting myself out of challenging or tricky situations. I really enjoyed the difficult ones, where you have to sit there and wait for them to come out. And while I was serving papers, I would go into the office every now and then to grab big printouts, and while I was there, I'd talk with Patrick, the owner of Lone Star Attorney Service. Every now and then he'd ask me for just random business advice, given my previous background. Over the next two years, while I was serving documents, we became good friends — I became an advisor to him, and we started trusting each other. And at some point, I believe it was last April, he was looking to exit out of the operator role in the business and step into an operator role in a different business. So he approached me, we had lunch, and he asked me if I wanted to run the business. I politely declined that offer and asked him if I could just buy it from him. And so we came to terms — great financing terms — and now I'm the majority owner of the business.

Collin McKee

That's awesome. So the buying process itself — you've got a great track record: West Point discipline, the Rice MBA, the Citi background. But you said even with all that, it still wasn't an easy process. Why was that?

Raj Kankaria

At an institution like Citi, whenever we're doing transactions for large public companies, the processes and what you expect are very specific and structured. Here, you have to go out and find a lawyer. When you're working for a company like Citi or any large investment bank, they have a team of lawyers that they already work with. You have to go find a guy to do some financial diligence — over there, they have an entire team in-house to do all of that. You have to trust your gut here: you have to diligence the company yourself and make sure there are no hidden skeletons. Whereas at an institution like Citi, for public companies, most of the information is fully disclosed. So there are a lot of unknowns here that aren't the case over there. And then there are personal risks that you take in these small business deals that you don't have to take when you're working for somebody else. I think that's what requires the largest tolerance for risk and intestinal fortitude, if someone does want to do something like this.

Collin McKee

So you've talked about this whole story as an "every door that closes, another one opens" kind of story. Does that framing still feel true when you look back at it, or has it evolved at all?

Raj Kankaria

I think it still holds true. When the door of my marriage closed, I thought it was the end of the world — I was never going to find a girl like that again, and I had dug my heels into the identity of being a married man and a dad. Two years later, I'm now with an amazing woman who has been a great partner to me. Same thing with my professional life. I thought that after I got laid off, that was the end — it was the worst day of my life, I'm never going to recover from this — and now I'm doing something I love way more than what I went to school for.

Collin McKee

Excellent. A lot of people probably would have walked away — had that horrible experience and not wanted to have anything to do with that industry, because it would probably scar them for life. Instead, you decided you were going to go head first into it and purchase the business. So before we get into the business itself, I want to understand what you actually built — what you took over once you got the keys, and what you've built. Let's talk about what happens inside the business day by day. You've drawn a pretty clear line that the service of process itself is not touched by AI — that's still human. Of course, you still need to physically serve someone the papers. But everything behind it is kind of different. So what did you do underneath the hood with the business that you're running now?

Raj Kankaria

Yeah, so when I bought it, it was about a million dollars in revenue. A lot of the processes were completely manual — there was a human typing in stuff on a keyboard, and every process, I would say, was touched by a human being. Slowly over time, we've begun automating. The first step was to automate the emails. We get hundreds, if not thousands, of emails daily asking for status updates, whether we can do service in a certain area, or job intake. So what we've done is use AI to help us automate that email management process. Now, there is not necessarily an AI solution for every process issue, and this was one of the ones where we realized that. We didn't necessarily need AI — we didn't need to pull Claude APIs or whatever to resolve this issue. All we needed to do was set up a bunch of different filters. But we did that with the help of the AI, of Claude. Claude would go in, audit all the emails, and it laid out the process and all the filters that we needed for the different mailboxes that we have. Then we have attempt management. There are some process servers that still send their attempts in via email rather than doing it via our software platform. Using AI, we were able to parse through the emails based on what their labels are, and we were able to figure out a way to take those emails and actually update the specific job in ServeManager, which we use — fully automated. So it goes through all the emails a few times a day, takes those attempts, puts them into ServeManager, and sends out the alert to the client. And there were several other processes like that that were happening with a lot of human intervention that we've now automated — up to probably 75%. Twenty-five percent of the jobs still require some human intervention, but we're about 75% automated. Billing and invoicing is another one. We used to have to post every single payment manually — we've now got an AI to be able to do that.

Collin McKee

That's excellent. Getting a lot of your time back, and doing the part of it that gets you that dopamine hit that you love. So that's great. You mentioned something in our first discussion that I wanted to dig into. You said when you automate too much, like one out of three or four jobs tends to break because it needs customization, and that actually creates more work for you than if you'd never automated it. So what did that teach you about where the line should be, as far as where you use AI or automation and where you don't?

Raj Kankaria

I think AI needs to be used on simple tasks — simple, repetitive tasks. Where AI breaks is where customer service or customizations are required. So for example, I'd say seventy-five percent of our jobs are probably routine, run-of-the-mill: a document is sent to us, the address is correct, everything is right, we send it out for service, it hits the right address, the guy gets served. But then there are the fringe jobs, where you have to coordinate with the server because the client wants it done ASAP at a certain time, and it might not be the right address — and if that address doesn't work, then you have to go to a different one, and there's some hitch along the way that requires some more human intervention. So it's good to have a human on backup, but then to scale those simple jobs, because they're automated. And so I think that if we can maintain that ratio of those simple jobs plus a little bit of the custom jobs, then we can scale with the processes that we currently have and charge a higher margin on the jobs that require more customization.

Collin McKee

Smart. And there's another distinction that you made in our first discussion that I think is lost on a lot of people, and I wanted to kind of unpack it as well. You said automation isn't the same thing as AI, and throwing AI at a broken process doesn't fix it. I love that. Would you expound upon that and what you mean by it?

Raj Kankaria

Yeah. So we live in a world now where many people think that throwing AI at a problem or a process is what's going to solve the issue. But what we found is, when we did the audit of all of our processes, there were some processes that were just broken. Whether you automated it or whether it was done by human intervention, it was just a broken process, right? So for example, our collections process. The way that we were doing it — even if we automated it, even if AI did the exact same things that a human was doing, it would not work. And so it required more than just AI intervention. It required really going back, looking at the process, fixing what needs to be fixed, and then automating the parts that could be automated, and then using AI at certain steps of that process — rather than just saying, "Hey, here's AI, why don't you go and do what the humans were doing in this broken process?" And I think that many companies are taking that approach, where they're just replacing humans with AI but not really examining what the flaws in their processes are.

Collin McKee

Yep. You've got to look at where your foundation is. If your current state of how you do things is broken, throwing fuel at it is not going to be the answer.

Raj Kankaria

Exactly. So if your customer service is crap and you replace humans with AI, it's not going to solve your customer retention issue, right? The AI only works as good as the humans that train the AI. And if you can't train a human, how are you going to train a robot?

Collin McKee

Exactly. Well put. Another thing — your team seems to be kind of young. You could say inexperienced, but in this day and age, the younger ones are actually more experienced when it comes to technology a lot of the time. You've got essentially high school graduates that are using AI daily at a level that's higher than other large professional institutions in your space. So what do you think changed for them? How has AI allowed these inexperienced workers to work at a level that's so high?

Raj Kankaria

Yeah, so there are some interesting statistics that I've looked up. Apparently, less than twelve percent of the US population currently uses AI on a daily basis. Not just for work, but generally speaking — for work, for life, personal, professional, whatever — less than twelve percent of people are using it daily, and that's both tech and non-tech roles. When you look at people in tech roles, you have thirty percent of the population using AI on a daily basis. We've developed a culture where, as soon as we start up our workstations, we open up Claude. And what I've seen is an increase in critical thinking, an increase in productivity. I'm seeing things move a lot faster, problems get solved faster — and just a culture in which we have fully adopted the use of AI on more than just a daily basis.

Collin McKee

That's awesome. That second stat — the first stat I don't think would shock people, twelve percent. But then those that you would assume are using this new wonderful tool — a third. A third. Amazing. So in that same sense, you work with a large number of law firms — not just there in Houston and Harris County, but those that have dealings in Texas as well. What's your honest read as far as the median or mean usage of AI across the law firms that are clients of yours?

Raj Kankaria

It's pretty low. I mean, it's so low that I can't even put a number on it. I don't even know what they're using it for. I think the legal industry — what we saw was a huge increase in the use of AI, and it was like the Wild West. They were using it for anything and everything, right? No real purpose — just, if it's something that makes a human being work, try to replace it with AI. And so what we saw is AI started drafting petitions, and AI started doing legal thinking and legal work and drafting these petitions with false information. So it was putting in precedent — precedent law that didn't exist — or coming up with stuff that was just hallucinated, I guess. And so we then saw the legal industry take a step back from AI and think about where it could be used, and I think right now everyone's just kind of taking a pause from it, almost, it feels like. But I think where the potential lies in the usage of AI in the legal industry is not necessarily in the practice of law, but rather in a lot of the front end and the back end. So intake — I'm seeing a lot of AI being used. We don't really interface with that, but what I've seen is intake being largely either automated or outsourced. So whenever you call a law firm with whatever issue you may have, whether it be personal injury or some other kind of litigation, the information is being evaluated by some kind of LLM, right? And then the phones are being answered by someone overseas, even, sometimes. Then another place where I see potential for it — again, this is not somewhere where we interface — I believe that court reporting and stenography are largely going to get disrupted by AI. Like AI transcribers — we see it on Zoom already: Fireflies, whatever you want to call it. That is going to just devastate the current court reporting industry. Then the next place where I see it is in a lot of the litigation financing. So creating algorithmic models for evaluating lawsuits and underwriting litigation funding — that's another place where I see a large potential use for AI. Where I don't see AI being disruptive at all is in the actual work that lawyers do. So drafting these petitions, coming up with case precedents, arguing in court — I don't really see much disruption there. And then another place that I see it, too, is jury selection. I think there's a lot of potential there. So I think in the ecosystem surrounding litigation, there are a lot of potential uses for AI, but I think there's still a degree of evaluation that needs to be done on where it's going to be the most useful without impacting the attorney-client relationship.

Collin McKee

Yep. It goes back to — it's not going to do the core of the business, what they got into being an attorney for, the thing they're passionate about, their dopamine hit, right? That element can't be replaced — the experience, the personal experience, that a lot of these attorneys have. Nothing to worry about there. But as far as the monotonous, repetitive work — yep.

Raj Kankaria

Yeah. And the ecosystem surrounding the legal industry, I think, has a lot of room for AI disruption.

Collin McKee

So if you're talking to a managing partner right now, what's one of the things on the vendor side of their business — companies like yours — that you wish firms understood more or thought about more that maybe they don't today?

Raj Kankaria

I think that they see these legal service providers as commoditized businesses — who's going to give me the best deal, right? And what I ask them to consider is: are they getting the highest quality per dollar spent? There are firms out there that are just complete crap, charging the lowest prices but also giving crap service, and then there are companies out there that are super white-glove, and it doesn't really fit the needs of the firm, right? And so I ask these attorneys — actually, our audience is mostly paralegals — to really think about whether that vendor fits the needs of the firm, rather than viewing the vendors as a commoditized business, like a gas station, right? It doesn't matter what the gas is — people just go there based on price. A lot of people see businesses like mine as just commoditized businesses, but that's not the case, because it's a business where speed, accuracy, and transparency are critical, and not everyone provides that.

Collin McKee

So say I'm a firm right now, and here I am, nodding my head, agreeing with you — but that doesn't mean I'm going to action anything. What's the first thing that a firm would actually do to evaluate the relationship that they have with their vendor, or these back-end processes? How do they make the right decision and make sure that they're getting the best value for their money?

Raj Kankaria

Yeah, I mean, the best thing to do is just try us out, right? It's pretty noncommittal — we don't ask them to sign a contract or anything like that. Just try us out. Send us a job or two, see how we perform, see if you like us. If you don't, I have a dozen people I could recommend you to. If you like us, we'll continue working together.

Collin McKee

That's fair. Very fair. So, quick rapid fire before we close out the episode. What's one thing you think a law firm should audit about their vendor tech stack, say, this week, if they have the opportunity to?

Raj Kankaria

Man, that's a tough question. I'd say probably just how many vendors they are using currently, and where there's room for consolidation. So, are you using a bunch of different process servers for a bunch of different counties that you're operating in, or do you have one person handling all of them? Same for court reporting and stenography, same for all the other litigation support services. Do you really need all the vendors that you have, or can a lot of them be consolidated into one platform, or a few platforms, so you don't have to keep track of a bunch of different service providers?

Collin McKee

And that kind of bleeds into my next question. Where do you think all of these process and back-end services go in, like, the next two years? Do you see a lot of consolidation? Is that where you think we'll be in a couple of years' time?

Raj Kankaria

Yeah, I mean, I think a lot of these litigation support businesses are highly fragmented. I think there is definitely room for a lot of consolidation. I think that what it's going to come down to, though, is: are the consolidators going to want to spend the money to acquire, or are they going to want to spend the money to advertise and organically grow and just let everyone else die out? Either way, there is going to be consolidation, though.

Collin McKee

Yeah, I tend to agree with you. And lastly, where can people find out about you and follow your work?

Raj Kankaria

They can visit our website, www.lonestarattysvc.com. They can follow me on LinkedIn. We have Instagram, we have a Facebook page. But yeah, I think the website and LinkedIn are probably the two best places.

Collin McKee

Awesome. Cool. We'll make sure to put all those links in the show notes. So Raj, I really appreciate it. This was exactly what I was hoping for. I appreciate you sharing your story — it's probably one that some people would try to shy away from, but you own it, and you've come out with a fairy-tale ending there. You're back in a great place. And for anyone listening, if you're wondering where your own firm or your business actually stands when it comes to AI — whether you're exposed, if you're behind — that's exactly what we do here at Endeavors AI. We like to work with firms and small businesses and see how AI-ready they are. Some people are sticking their heads in the sand — that's not where we need to be — but we need to be disciplined and respectful of the large language models. They're here to stay. So reach out to Endeavors AI; we'd love to help you with that. But with that, I'm Collin McKee, for Raj Kankaria. We appreciate you watching, and we'll see you in the next episode.

Raj Kankaria

Thanks, Collin.