AI in M&A: Building, Buying, and the Future of Dealmaking // Kison Patel // #315
55:32 Share

So name, Kisan Patel, CEO and founder of Deal Room, host of the M&A Science Podcast, coffee. I like my oatmeal clates. People of Earth, we're back for another MLOps Community Podcast. I'm your host, Dimitrios. Today we're talking all M&A. We are also talking about the product lifecycle and the inception of where and how you can validate injecting AI into your already-

fairly mature product. Let's get into this conversation. I know you wanted to talk about the founder's journey that you've had. I am particularly interested in that right now because it feels like every once in a while you go through the slumps in the founder's journey and I have been or feel like I'm in a bit of a slump and so it's good to hear the rocky cut scene and how you're going and doing your training but

giving it to me a little bit more straight and helping me to understand that it's not just a five minute scene. It can last for days and weeks or months or years. So I'd love to hear just about like the inception of Deal Room and then how it's been so far. Yeah. So I come from an M&A advisory background. So I started from the bottom. Long story, I guess, behind the story, behind the story, which is

Failing out of undergrad, you know, found myself in real estate, trying to build a career to circumvent not having a college degree. And I ended up finding my way. I didn't find selling houses my jam. I just, it's like an emotional sale. And it wasn't, people don't want to buy houses from 21 year old kids. That's the thing. But somehow I ended up joining this startup where it was two guys and they were building like this business brokerage type of practice and

In hindsight now, I see that they were generating all these inbound leads online and they're mostly Indian descent people looking to buy like gas stations and hotels. And they probably figured, you know, we can barely understand it when I hire this kid and let him deal with them.

And that was it. It was like day one. I'm on the phones talking to prospective buyers, figuring out what they're looking to buy, contacting businesses that fit their profile, seeing if they wanted to buy those. If they didn't, can we list them for sale? And just got my firsthand experience of what M&A was really early on.

And it was funny because I had to go backwards. Like I would learn to put the deal together, but then I had to go learn how to do the analyst work. Like, how do you actually put financials together? How do you do, do all the modeling and things like that? Um, I did that. I left after one year cause I figured, Hey, I already know everything here and this firm doesn't really have a clear strategy. So I left and started my own practice. It's probably still pretty young, like 22, 23 and built that up over about nine years.

um ended up doing most of my deals in hospitality so i think at the peak was working with kimpton hotels extended state america la quinta and then i ended up finding fig like an opportunity to work with the financial institutions worked with a couple regional banks on some buy side and then sell side for small community banks that led up to the recession in 2007 when everything got flipped upside down i wanted to do something in tech and i uh

I had this idea of like buying these financial domains. And what I ended up doing was basically creating these automated blogs that would populate content based on the keyword of the domain. And I remember at the peak, I had like 500 of these finance keyword domains because I noticed the ad revenue was higher than average, like anywhere from five to 10 X higher than a typical blog.

A web domain. And so we ended up with about 200 blogs that were powered up and we're building like a customized platform to power all these blogs. I mean, I didn't realize at the time, like we're basically gamifying Google's algorithm to rank all these sites really high. Free LLM days.

and then one day they just uh made their penguin update and considered all our uh linking approaches back as link farming and then it blocked everything so our climbing revenues went to zero and i had a small team of about five folks we had to basically wind down the whole business and um

The one thing from that experience was the first time I was working with software engineers, I was really intrigued how they were using Jira to manage developing our software platform.

And that led to this inspiration of why not build a project management tool for M&A. That's what got the idea to start Dealroom in 2012. So JIRA for M&A because every software developer loves JIRA. That's probably a better way to frame it. I should probably use that. Yeah. Think what you do. But the fascinating thing here that I really like is taking something that could be done in a spreadsheet

Right. Or that is traditionally done in a spreadsheet and just giving it something beautiful and helping folks along on their journey, as opposed to you need to have this knowledge and figure it out from the spreadsheet.

That's true. I mean, if you look fast forward from that time to now, the product has evolved. We differentiate ourselves that there's a lot of tools to sell businesses, but we are very much focused on the buy side. So if you look at our customers, they're large Fortune 500 companies or private equity backed roll ups that are consolidating some and you see them across the board.

You see like HVAC, service businesses, wealth management, accounting, private schools, you name it, managed service providers. There's just a lot of consolidation with those smaller fragmented industries. So those, and then the product itself is like full end to end. Now it's managing your pipeline, managing your due diligence, all the back and forth requests for information, clarification questions, doing your integration planning,

actually executing your integration, any kind of like BI reporting you want to do. And now more recently, we started building some AI capabilities for extracting data out of contracts. Yeah, I want to get into that because it feels like a perfect fit. But the part of me that I want to make sure we cover is in a deal, in the deal lifecycle that you just laid out,

you're not having the marketplace for the companies that are for sale, right? And this is, and you said you're focusing on the buy side. So I imagine these folks are trying to mine for gold out there in companies or small businesses that are willing to sell. We're not, it's funny you mentioned that we're not a marketplace, but, you know, and I feel like this is the

I feel like founders now should probably know better because there's so much content written about this. But back when I started, there was nothing, there was no startup ecosystem anywhere, especially originally it was in Chicago when we started the business. And I,

At the time, it's like you start listing out your features that you want to develop. And this is like the definition of feature creep, where you kind of had this idea to build a project management tool for managing M&A. And then you end up with a list of like 100 potential features you could build.

And then you're like, where do I start? And then we have the idea of starting with more of this marketplace model because we're like, we should probably start with connecting buyers and sellers. And then as they progress with the process, we'll build those features out as needed.

Our MVV should be, you know, in the marketplace. And we did. Within six months, we got our initial launch for a marketplace. We got about 1,300 users on there and 200 deals. And I looked at this and I was going through the deals and realized we just built the world's most sophisticated dumpster for deals because everything on there was just garbage. It was like nothing a person would really seriously consider investing in.

And that's where we had to like go back to the drawing board. And I think around this time too is when all like the lean startup was, was taken off. Well, there was a book called running lean. That was pretty cool because it was more of a practical approach on how to, you know, leverage the model. And so we started doing these customer development interviews and really,

Trying to be as unbiased as possible. I think that's like the number one threat to an entrepreneur is yourself. Is that you're like in love with your idea. You have this natural biases that you're completely blind to. I think there's new books now, like the mom test and things of like, how do you, you want to frame questions to people in a way that's unbiased. I think that's a whole approach and discipline of its own. Yeah.

It's a skill. And then it's just doing that, like doing the reps. Cause there's so many entrepreneurs skip it. They just want to hurry up and start building something, get to market fast. But I honestly think this is the most valuable step in the entrepreneurial journey is validating that you're solving the right problem. And I think we've done this and we continue doing this, right? This is like now in our R and D it's instilled. We have any new ideas of what we're trying to build for. We go through the same process and I've seen it where I've

If you do this right, you can actually find more interesting problems to solve than what you originally set out to do. Oh, wow. Because if you just go around to people and you're like, okay, this is a specific cohort of people that I think has this problem. I want to go solve it. But if you remove that biases and you're literally going to people and say, hey, what's the biggest challenge you currently have doing X? You know, whatever it is. What's the biggest challenge you have hosting a podcast? And then just let people know.

speak up and then you go through a series of those conversations and look for patterns um but that that's i think if you do it right you can more times than not actually find some more interesting problems to solve then you go through the whole exercise like okay you know one do you do try to steer them to the problem that you're going to to address uh because if that doesn't happen naturally that's one clue um

And then when you do start bringing up and listing these problems, how do you weight them? Is it a paper cut problem or is this like a severed, you know, hemorrhaging thing you really got to solve? That's what I was going to ask. Because that's where I tend to struggle is I'm...

full of ideas and full of cool things that I want to build. But then at the end of the day, figuring out which one is going to be the one that I focus on for the next three to six months or rest of my life, depending on how successful it is, right? That that's where I will get caught up because I

It's like in that analysis paralysis of is this actually going to be worth it if I put my time into it and have a few friends that have told me like, well, look, man, you can quickly test, try and get signals and then see if there's something there. And if not, it's not like you are that invested in it. So go on to the next one. Yeah, I would agree with that. The testing part's really important.

And I think just the conversations I found the most valuable. I feel like some people are afraid to talk to people and they end up doing these highly quantitative approaches and they send surveys out and things like that. And it's like terrible. You get a weird bias there because people that take the time to answer those already...

Yeah, at a different level. Or even I've seen folks that will use LLMs as their way to do this type of research. And it's cool for some, but it's definitely not...

equal to getting on a call with a prospective client. It could definitely help you guide you where it's like we do it. I'll use it to key up questions. I'm doing one right now for our next set of AI features and I'm using AI to help really generate the question flow and especially it'll give you too many and then you got to go back and say, all right, let's really prioritize this because I want to get down to about three or five key questions because the goal is to make it conversational.

Like that's where you get more of the interesting insights is it's conversational because you want to peel back layers. And if you just kind of like tactically ask question after question after question, one, it sucks for the other person, but then it's like too high level, um,

Uh, the interesting insights are sort of like when you start getting some real understanding, like, why is that? Like, why is that? That, that kind of the cliche thing, right? You ask why five times and you really get to some, some clear root cause of all this stuff. That's more of the important thing to, to have in those conversations. Um, I, you know, I also think that when you reach out to people, cause you, if you can reach out to people cold and you're like, Hey, I'm trying to solve for this problem, um,

And I wanted to get your feedback, you know, your experience and perspective would be really helpful. Just the response rate is probably an early indicator if you're solving something interesting or not. You know, you reach out to a hundred people and nobody responded. It's like, okay, maybe there's something there. You know, rethink that, try to see if we can amplify it a little bit and see. But, you know, you should be able to get some conversations. I think that's like the most valuable thing.

It gets skipped all the time or just taken lightly. You have like five or six conversations. I think a better target is like 40 because then you can really...

identify some patterns, but also cover multiple cohorts of different prospects that you would tap into. And you might find like the way the original customer base you set out for isn't the right one or isn't the one that really experiences the pain that you thought they did. I love that we're harping on this because it is so important. And I had a friend who has now created a fairly successful startup. They've been ramping like crazy. They've been

And he told me about how they came or stumbled upon the idea that they have. And they were talking to CROs and they went in with one idea. But in one of the calls, the first thing that they asked was, so what's top of mind for you right now? Or what's like the burning thing that keeps you up at night? One of those kind of generic, open-ended questions that you can ask. And the CRO said nothing.

one answer and they were digging in, digging in. And at the end of it, it had nothing to do with the product that they wanted to pitch them. But they said, hey, look, it's funny you should say that because we're actually working on something like that. Can we get together in two weeks and show you our demo of the product? And so in the next two weeks, they just didn't sleep and coded up an MVP, showed them a

And then they closed their first deal right there on that call. And they realized, wow, we got something here. And so then they just kept doing it and saw that, okay, this is A, really easy to sell if we can show them the value here. And B, we need to start hiring people. That's great. I love it. I call it being dumb. Because again, you just...

It's like our human nature. You get into a meeting and you're constantly thinking about the next thing to say, where are you trying to steer the conversation? But if you just come in the flip it around, just be as like dumb, like get to the state where you're not assume whatever you know is wrong or, you know, nothing.

And then you just like intentfully really listen. And then all you had is that one lead question, like what's on top of mind? What's the biggest thing you're struggling with right now? And you just really dig in and listen and understand that problem. That's going to steer you in a much better direction than having so many different things mapped out and where you plan to take the conversation. I was listening to a podcast with...

Gilly, who created the incubator that Wiz came out of, Wiz and a bunch of these cybersecurity startups. And he said that one of the questions that they have their startup members ask the potential clients is to basically to cut through the noise, because what he was saying, if I'm remembering correctly, is that

When you ask someone what their biggest problem is, it's very time-consuming.

specific. So this week, I have a very big problem. Next week, if you ask me that same question, I have a different very big problem. And so it all depends on when you catch me and what mood I'm in, what I've been thinking about, what meeting I just had, etc, etc. And so he was saying there were a few questions that he likes to have his startup founders ask, which is something along the lines of,

if you could rip out any one piece of technology right now, which one would you rip out? And the other piece was if you had to lose one piece of technology, but you would cry because you had to lose it, which piece of technology would that be? And so it's kind of getting outside of the box of, Hey, what, what's on your mind? Yeah. I actually like that approach a lot. Um,

I don't know why. I feel like both of them would be like Gmail. It's a plot twist. The same answer to the same question or the different questions. The biggest pain and the biggest joy. Yeah. So, I mean, I want to talk a little bit about how you've been creating these AI features, because I think that flows in perfectly here. And it is interesting.

very related to going out and I'm sure you went and talked to folks to try and figure out, okay, out of the million ways that we could put AI into our product, how and what's going to serve our customers the best. Yeah. So I started this journey last year and I got to tell you, like there's big thing I learned is, um,

When you have a company that's more mature, because we're about 50 people in the company now, that same playbook of when you're a company of like two to five employees

Does not work the same way. In fact, I'm still learning that. If anybody listening has some tips for me, because that's one thing that'll unfold in this sort of journey. But the beginning part starts off really similar. I think at the time I had about four different hypotheses on where AI would play well when it comes to mergers and acquisitions, specifically on the buy side when you go acquire a company.

And I remember starting from talking to existing customers and saying, hey, we got this AI stuff. And I remember I talked to a handful outside of it as well because you can get access to prospective customers and wanted their take as well. And what ended up was we quantified the value was actually –

What drew people in was the use cases where there was really clear value. And the number one thing that we found was people wanted to save legal expenses. So we kind of went through all these ideas and we ended up really focusing on contract analysis.

Um, specifically we started with this extraction that if you're going to go buy a company, every single company has a ton of employment contracts. They have a ton of customer contracts and they have a ton of vendor contracts and to sit there and go through them and mine for information. Cause like, for example, the first thing, if I'm looking at a deal, I want to know it's in their customer agreements. If there's a change of control provision.

Right. If we're going to buy this company and they have 500 customers and I find out every single one of them has a right to terminate the contract, that's a huge risk item. And I'm going to have to go to the seller and say, all right, before we close, like we need to address that. Yeah. Because I can't buy a company where I don't know if we're going to have customers who they all have the right to cancel their contract.

Now, normally you'd have somebody like a junior legal person that would sit there and they would open up each document and do a control F and look for change of control. And they just keep doing that over and over. And you spend quite a bit of money just doing that, that extraction. And usually there's some handful of key things you're trying to get in all those sets of contracts. You know, you might have a set of leases and you want to know what the assignment provisions are. Consent comes up in a lot of agreements. You know, you have to notify people about the transaction happening.

And this AI just does such a good job of, you know, taking basically unstructured data and structuring it and to be able to do that on a whole set or series of these very similar documents and sort of identify what's unique about it and come back to you and give you an analysis that explains all the different things that picked up. That we found just resonated so well.

And I think that was why we gravitated towards it. It's sort of the value of solving that problem. It was very easy to correlate to an invoice. Did you have a preconceived notion of, oh, well, normally lawyer fees or legal fees can be up to $1,000?

5% of the deal or X percent or 25K, 100K, depending on how big the deal is. No, like I was not thinking about that at all. I'm like the founder of Visionary thinking of stuff that'd be cool.

Like, it would be so cool if AI could automate, you know, this part of the deal. If AI could help with this sort of planning stuff and, you know, help you. Nobody cared about that. All this stuff, right? You think of cool stuff. But then when you sit there and talk to these people, you're like, wait a minute. They just care about saving a buck here and there. And if you, like, go back at everything, you're like, well, this would actually save you the most money. Yeah.

That's where we ended up putting our focus. And it was a good bet. Like that bet worked out really well. Like it was, it's great capability. But now if I look at this year, so this kind of goes back to dealing with a mature team. Because there's a point when you move through that process and you realize you can't do things the same way you did before. Like just prototyping and, and,

Code typing you can still do because you can work with a designer and get some mock-ups and get feedback, right? You should do that. But there's like a mature team. So it kind of gets, it's got to get handed off back to them and it kind of goes through their whole cycle. You know, before you might not have, you might not even have a Q&A cycle. Now you got like all these different steps that they're there and then you'll find yourself like breaking more things than that.

It has to go on Jira. What's that? It has to get put on Jira. Yeah. So I got it. That was the point I realized, like, okay, I got to step back and I'll step back and there's other things to work on in the business. So I, you know, but I think this is an interesting thing. I stepped back and six months or so goes by and I realized like there wasn't a lot of progress in that initial feature set we created. People use it.

But it just didn't mature or evolve. When you look at AI in general, it's evolved really quickly. So that got me thinking, when I looked at it and the way engineering resources were allocated, you're competing with a lot of things. You're competing with a lot of requests that your existing customers have. Your sales team has requests that they identify prospective customers have or what would help them sell.

And then I looked at it, it was like basically like 10% of it was only getting allocated to like real innovative parts of the product. And it's kind of an interesting thing I've noticed. Like you could be this new AI startup and you're just like 100% focused on it.

Which is they have their own set of challenges, right? Because right now the market's getting flooded with AI tools. This getting distribution is really tough because you're especially in like finance industry. People are pretty conservative. They haven't heard of you. There's that part. There's a security they're really sensitive of. Have you figured that part out? Just the whole nature of working with larger companies and things like that.

Those are like challenges the smaller companies have. Then the bigger companies actually have their own challenges because it's all competing priorities. Like that's the whole dilemma, like how you really innovate well. So we actually we have this smaller subproduct, but we ended up that didn't have a roadmap. And that was like growth of slowing. I said, hey, when I do a founder reboot on that product, I should move the AI roadmap onto this subproduct.

and then i started like just building a small little team to really just focus purely on ai development um and just you know we're able to accelerate so much faster by doing that because now we didn't have all these conflicting priorities yeah um you gotta and then we also like got a lot of the benefits of the maturity right we already have this mature platform we already have security credentials all over the place all those things are figured out we already have some reputation in the market there's already a base of customers

So it seemed like that was sort of finding that little area to get that right balance was a good thing that now allows us to really accelerate this roadmap. How's the speed of iteration been since you did that? And if I'm understanding it correctly, it's like you created a sub product under the same brand. So you're not doing anything. So it is a different brand, actually. It's called. Oh, is it?

Yeah. You know, the backstory on it real quick is basically we would turn away all these like small little companies and we realized like all they wanted was a data room. They just want data security product, right? There's Dropbox and things like that. But when you're doing M&A, you want to have audit trails, you want to watermark documents, have very granular permissions. Those are like the differentiators of a data room. Like we technically have the data room in our product. We had the idea of carving it out and making it into a self-service product. And we did that like five,

five years ago and didn't do anything no marketing it was just you know hey if you're our main product wasn't a fit we'll send you to our sub product and that thing grew it grew to like a million and a half revenue so we're like this thing actually is doing extremely well for what little effort that we put in because we didn't do any like we don't we put the stood the thing up we never did any product updates or anything forgot about it yeah and then you realized that well

You're going to take that and create the AI features inside of that product. Let's let that be our R&D and kind of fork off like a little separate team. Now, it's the same back in engineers. So it still rolls, you know, parts will still roll into the same engineering team. But they got their own free will on how they want to allocate, you know, who's working on what.

but everything else is completely separate it's like its own uh you know product marketing design all those components are just a small little team which i much prefer

Yeah. Yeah. I think that's like a whole founder frustration of watching your baby grow. You know, people always talk about baby grow, but like baby grows and just get slow as shit. Like everything takes forever to get done. And that's to me the, the hardest thing to, to accept. When you're used to moving at the speed of a two, three, five person team, and then you go to a 50 person team, I can only imagine. And then you're

If you go to a 500 person team or a 5,000 person team, it's got to be so painful. I just, I don't know. I mean, it worked out well when we brought in a COO. When we brought in a professional person to run operations, that was God sent for me. Like it allowed me to focus on what I wanted to in the business and have somebody that really enjoyed doing all the stuff I hate doing.

which is managing people and, you know, the operational components of it. So I think that's, but you still like, I don't know, I can still think of pulling my hair out at how slow things move and how much more inefficient it becomes as you grow. And it's just, there's, you want to be proactive to counter it, but you just can't, it's inevitable. Like it just,

Who knows? Maybe with AI, we're going to figure out AI agents that manage other agents that will do everything. And maybe that'll be the counter move. Yeah, that's the dream. I am very skeptical of that happening anytime soon. And I see all these posts that folks are talking about, like, look at these companies that got to 100 million in revenue with only 10 people or 30 people, et cetera, et cetera. And what I see as a commonality between those companies that

people do not highlight in their posts is all of these companies are product-led companies and you can just give them a credit card. And so, yeah, they have 30 engineers that are just working on the product and then they have some virality and they don't need to necessarily hire a big sales team or hire the big marketing team. And so I've got a little bone to pick. I would say that's my only counter

argument on that totally agree I mean there's some lottery winners out there right you just struck the right chord and things lined up for you there's also this uh easy come easy goes because there's some that kind of hack their way to to get to the early traction um

Um, but the reality is of all the companies I looked at, they are struggling. They are really struggling with the, you know, I, I feel like there's some benefits of AI allowing you to get to market quickly. Um, but there, there's still like a lot of stuff that just,

I don't know how you can accelerate it. You know, there's building a brand. And I think that's the thing I still would rethink if I were to do it again. Like for us, the podcast and social media is now driving about 50% of our new business. And if I were to, I mean, I'm staying anchored in this industry because we have such a good foothold with those channels. But if I was to do it again or start from scratch a new industry, I would focus more on how are we going to solve for distribution? Yeah.

It's one thing when you got followers listening to you that helps quite a bit. But if you're entering a new area, I think that that's a bigger challenge is the distribution over the product. And that's becoming really well known. That's why they have that saying, right? The first time founders focus on product and second time founders focus on distribution. Yeah.

That's exactly, I'm case in point. I would focus on distribution. Exactly. So I want to talk to you about pricing because I feel like that's another interesting piece. When folks are creating AI products, especially inside of a product that they already have,

How do you look at pricing? Is it usage based? Is it just like, hey, here's this AI feature that you can pay X amount per month for. And then you get access to that feature, which later on your side, depending how much they use it, it can be a really good profit margin or a really bad profit margin. Yeah, that's a good question. One, I don't have like a really simple answer at all.

I can tell you we take a mixed approach of just looking at what's trending. Because there is like some truth that if there is some established status quo to pricing, now AI is really emerging, so a lot of areas it isn't. But if there is, sometimes you really don't need to like disrupt the pricing model because it's

things could be priced really well. And if you're sort of, everybody's kind of price similar, like that could be a benefit and you just selling differentiators, you know, you don't have to sell yourself short. So, Hey, if there is some pricing model out there and it's like pretty aggressive and it's what standardized out there, I wouldn't mess with that. I would adopt it and then just focus on how do you differentiate against competitors? That would be a better win because otherwise you don't want to come in as the race to the bottom or you're trying to, you know,

you know, be the most cost-effective solution unless there's a real strategy behind that. Like you're aggressively going for market share because you've got that tailwind of acquisitions at a low cost. The other thing I think that we're realizing is

When you have a certain pricing model, it makes it more complex. If you have a pricing model that's unique, like in the data room, you tend to charge by data. It complicates things because if you're looking at just AI products and that's all they're billing for is this particular AI service, when you have a core product that's already existing with a certain pricing model,

You can't just slap it in another, say, hey, we had this new ad capability and we're just going to charge for it because they charge for it that way. That exponentially complicates your pricing. Exactly. So I think that's like where our struggle comes in. It's like, whoa, how do you figure that? So now for us, it's like we're actually looking to simplify it where it fits in the same pricing model, where it's a bit more modulized, right? You're paying for a bucket of data for X amount per month.

You're paying for, what do we have, additional services. You know, you may have extra rooms that you pay X amount for monthly. You know, you may have some tier for better support you're paying X amount monthly for. Now, like, AI capability becomes that. It's more of like a flat monthly fee. Yeah.

But then, yeah, your engineers will come raise a concern. Hey, by the way, if somebody runs this, you know, prompt on this 300-page contract here, it's going to cost, like, so much dollars. And if they keep doing that for so many of them, you know, we might lose money on this. And that's where it gets interesting because...

People don't like the complicated things. And I've noticed that with some of the companies where they've tried to do the per credit or per consumption based. I think it works at some level with like middleware, right? Or Amazon, we're just paying for our usage.

But some where you might be dealing directly with end user, they don't like it. They're used to their chat GPT, which lets them prompt all day. It's 20 bucks a month or 200, depending on what tier you have. Yeah, that's what I'm always thinking about because you bring up a really good point on the complexity. And then sometimes I imagine if you're trying to sell this for the deal room, for example, this product-led product,

product. And so it's, I imagine it's just you swipe a credit card or you pick your tier and then boom, you're in. If all of a sudden you're like, well, if you have two terabytes of data, then if you want that AI feature, it's going to be this much, or we'll give you X amount of credits. And we assuming that you're going to use it this much. So that should be, you know, it just is like, wow, can I just have a simple solution that

that I say, all right, it's an extra 10, 20 bucks. Cool. For this AI feature, I need it. Let's try it or maybe not. And so that's why I've heard a lot of folks talk about how in iteration stage or in that product validation stage, it's very common and it's almost like best practice these days to just outsource everything to a model provider API.

Because you don't have to worry about any of the infrastructure. You don't have to worry about any of the headaches of serving these models and all of that. Once you've validated it, though, and you see people are using it and you see how much it's costing you, then you can say, all right, we should probably bring this in-house. Or we should at least look at how we can drastically reduce costs. Can we use the smaller model and still get similar results? Can we try to do things that are

not necessarily just pinging the API. And when you were describing that, I was thinking, because we have two products, right? Deal room is sales-led and then firm room is product-led. And you think about pricing so differently between those two products.

because uh sales led it's literally sales led like go let the sellers figure that out and they they will they'll come back and they just try to sell different ways and they will sell for a much higher rate you know they can sell ai on credit they can do all kinds of things they can sell it and that's what they do is they actually have um they end up with tiers of credit with so many whatever it is i don't know if it's either documents i think it's by the document um

And so they bucket it that way. And then they do get a good price for it. But then when it's self-service, the same model doesn't work at all because, yeah, you're in a good sticker shock. But, and two, people, you know, the way we made it work is just like very transparent, very simple pricing. I think that's the frustration point we notice in the industry. People, all our competitors have like all these hidden fees and things like that. And that's actually where they make a lot of their money off of.

So we're trying to, as part of it, you know, how do we serve the customer best? And that was eliminating the surprises. I've seen the other common thing that I imagine you've probably thought about or you are currently doing is

just giving folks the product for seven day free trial. Hey, you've got Slack AI now in your Slack workspace for the next 14 days to try it out. You can do these things, blah, blah, blah. And then seeing if they want to continue, it's like, all right, cool. Upgrade to this plan. It's 20 extra bucks a month or it's 50 extra bucks a month. You know, it's been interesting. I haven't seen that work a lot of times. Yeah.

Yeah, that's why I'm asking you because I personally, when I get that, I'm just like, get the fuck out of here, AI. What are you doing? Like, I don't want this right now. If I want it, I'm going to go and find a way to use it, but not necessarily like shoehorning it into random products that I wouldn't really associate with AI giving me that much of a extra boost. So sometimes it's that, that it's like,

Like when I'm scrolling Instagram, I don't want to be thinking about work stuff. Isn't it funny how like our behavior, like our user behavior shifted? Because like I remember being, you know, like 15, 20 years ago, like I would try just new shiny things or try new shiny things because there wasn't a lot of new shiny things out there.

Now you fast forward a day with all the cutting edge AI, which you think would like, oh, wow, this is all game changer technology. But we're just so inundated with all these new shining things in our inbox. Like it's out of control that even from the brands that we trust,

that are pushing the new shiny things. Like we're still like, nah, like at this point, it's gotta be something that I, I want, like, I realize that there's this real, um, aha moment of like, I'm going to benefit and get value from that, which I think is like so much harder to reach nowadays. Yeah.

It goes back to what we were saying where you have to talk to folks and see what it is that they want so that you can hopefully get in front of somebody that also has that same pain. And they go, oh, cool. I was looking for something like this as opposed to a lot of times. And maybe it's the way that it is messaged or maybe it's the way that it is put in front of me. It's just like, I don't really care about this AI feature for some random app that I use.

Yeah, I think that's such a true thing. Like it is. The other thing that I wanted to ask you about was in the AI development journey of these new features for Dealroom and the other one is Firmroom. Firmroom. Okay, cool. The process of building a team. What did that look like? Because I'm fascinated by...

how a lot of ML engineers have now become AI engineers and how if you look at the job postings of AI engineers, it's basically like, oh, that's just a data scientist or, oh, that is a ML engineer, like a platform engineer. And all right, that is so AI engineer encompasses this whole world of AI

what these old job titles were. Maybe it's a full stack developer that knows how to prompt a little bit. Maybe it is, I saw one today where it was like machine learning systems and basically it was a data visualization job posting, but they were calling it this whole complex term. And so as you went and you said, all right, we want to build AI capabilities. What team did you assemble?

i work directly with our vp of engineering let him figure all that out because i i know that's what's going to happen it's going to become messier and more confusing and i'll take um design discipline as an example and again like when we started this company back in 2012

I mean, just to find a UI designer was a unique thing. Like that was extremely hard to find a UI designer. Somebody just designs interface. Yeah. And I remember that, like just going through all this effort to find somebody. And now if you look at like this design discipline, it's all over the place. Like there's just so many different facets of just the UX part, UI. It just, and it's still, um,

Yeah. Use the term product in there. And it's just a lot more fragmented and confusing. So you got to dig in and just figure out like what people, what do you actually need? Like really understand what are your specific requirements and then what are people really good at? I think that's like, I think that's the part of it too. It's like you got to still write your good posts.

to make it clear like what do you actually need because i i can see the fault of the the business itself too because we don't write good job descriptions you put stuff out there that's just you know you're asking for like every skill in the world that uh you expect the person to do but i i think if it's you sort of use more of that narrative of like

Pulling them into the problem you're trying to solve and what do you need? Like, what are the skills you need? And that's sort of really clear of like, what do you prioritize the most of? I think that's the thing I've noticed more recently when we had to hire the designer for our work. You know, I knew the thing we lacked on the past project when we used in-house resource was the UX side of it. Now that, you know, we had an opportunity to bring somebody new in, that was like the area we wanted to emphasize more.

But it was also like striking the right balance. Like, you know, we didn't want somebody that's just so strong on UX, but not so much on the actual UI design. That's tough too, because you get somebody like that and they're like maybe a little too much of a true UX architect that they just sit there in wireframes to lend a time. It reminds me of basically...

building any team, you're trying to figure out, all right, if it's a football team or if it is your engineering team or if it is a DevRel team, even you're figuring out, okay, what qualities are we looking for? And I really appreciate how you break that down and say, this is the goal that we're trying to accomplish where we had a bit of a hole.

or there was a vacuum, is in this specific spot. It's our wide receiver. It's our striker in football, depending on which side of the pond you're on. It's our content creator for the DevRel team, whatever it may be. And then you figure out that culture fit and you figure out like,

We have to make sure that they have a little bit of crossover so they can still be in the team and on the team and contribute back to the greater goal as opposed to just off on their own in this little igloo. You know, there's a lot when you look at how business grows around the culture component because you got to define it, especially the bigger you get. It's just got to be very well defined to attract the right talent. I think there's a lot of leadership stuff too that's

That's interesting to see how companies mature very differently. I think for us, like I struggled with it a lot in the beginning. So I started reading leadership books and found that they are actually helpful in learning leadership. I actually learned more from organizational psychology books. Oh, interesting. And you kind of understand like what makes for this high-performing work environment. I remember there's like three key elements. It was like a strong communication framework where everybody feels their voice is heard. Then two was communication.

acknowledging or recognizing achievements. They just take the time. People put in the effort they want. That's a huge thing that people want to feel valued for the work they contribute or the work is being valued. And then the third was like creating an environment where people feel their work amongst friends. You know, they're really comfortable to have like improv conversations with people and just reach out when they need to.

um those are the three things and then then you get into like all the values fun stuff of like how do you define that for the company and sort of like use that for as part of your your expectation of behavior and hiring and things like that that's um all fun stuff that matures uh as a company grows and then i for me it was always like empathy like i was not a very empathetic person i think some people they uh

May resonate with that. But that was like one thing I had to like really learn to get better at. And I think that was the thing that really contributed to when you have to manage people, connecting with them in that way. You're so soft-spoken. I would have thought it was the opposite. You were going to tell me that. I was like a very introverted, reserved, shy person. Like I would avoid confrontation and all these things. May contribute so to podcasting. Helps you with some of it.

Well, the last thing I wanted to ask you was about they say companies, great companies or good companies are bought, not sold. Have you seen that to be true now that you've had so many companies get bought, I imagine, or helped deals go through? I would say yes for one big reason. And I'm like recently did a whole series called Buyer Led M&A.

And basically the theme is if a buyer controls the process, they get better outcomes. But it all starts with defining your strategy that, hey, where you have a company strategy of where you want to be in 10 years. And if you look at an M&A strategy, what can you do or what businesses could you buy to accelerate that? You know, for us, it's at some point we want to get in Europe. We know just going to Europe from scratch is like a slog of a journey to do it unless you got the right leader that you can find, which is really hard.

But we do keep tabs on businesses that are already established with that similar type of customer that we have in those European markets. And for us, like that would be a good play for geographical expansion to help us accelerate that geography. So that's where you have that clearly defined. You approach those businesses, those businesses that are bought.

versus some banker or somebody comes to you and says, hey, here's this company in Europe for sale. And then you're trying to like rationalize why you should buy it. Which sometimes that happens. You found your entrepreneur, you like it, you're happy. You're like, hey, this will help us grow. But it isn't like clearly tied to the strategy. That's why it's so important to have that, you know, real clear business strategy. But then as you see a fit for M&A as a tool,

Uh, then you would have that defined M&A strategy and you'd have like a clearly defined criteria of the companies that would fit in there. Then you could be very proactive about like finding those companies and you get a better story to tell because you have like a clear understanding of why like better together makes sense.

Because these companies that are servicing those customers in Europe, like I know that we would create a lot more value because more times than not, like our ACV is a lot higher than what they're selling. They might be selling at like five, six thousand a year ACV and we're like 40K. Nice. So just to say, hey, look, like if we came together, you know, we could service your customers, but we could probably upsell a lot of them, like half of them to this premier product.

and be able to really generate a lot more revenue. We'd be able to support you. We're really good at this marketing side from what we've done that's helped contribute to our growth. Can we run some of those same playbooks to your business?

Um, that, that sort of like, you know, there's more you can get into, right? Like the, just the nature of doing a deal. Now all of a sudden it's like very amicable. I have this clear vision and strategy, how we're better together, but we work together and manage the timelines and make it for a really smooth transaction versus when a business is sold. It's like, here's this advisor that wants to sell the business as fast as possible. So you can go sell the next business as fast as possible and make as much money as they can at the end of the year. Yeah.

And so the timelines are super compressed. You don't get to do as much diligence. You didn't get to like think through of how you're going to integrate the tech as well as you should have. And then all of a sudden you buy the company, you go to integrate it. And it's like a lot of these thesis or hypothesis weren't what you thought they were. So I'm a big proponent of the bought oversold. Yeah. You talking about that reminds me of there's some book. It's like the

um failures the worst failures or 101 mistakes something like that and it talks about all these m&a disasters on how you have gigantic m&a situations that go nowhere and basically the companies these big companies that buy other big companies have to write it off to zero and you think about like the um

Who bought AOL? Yeah, a very clear one where, and no shade to anybody that's still using their AOL account, but I think we've moved on. That was one I remember being in there. And then there was a few others that it just shows you have an idea of what

can happen. And I think a lot of it is that, like you're saying, where someone will come to you and say, this company is for sale. We could do X, Y, Z. Wouldn't it be a great fit for your company? And then at the end of the day, it's not a good fit. And you got, I like that term deal happy, a little trigger happy on the deal. Deal fever. Yeah. Yeah. Deal fever, man. That is it. So is there anything else you wanted to talk about before we jump?

We got a good run. We kind of talked a little bit about the trials and tribulation of the founder journey. There's, you know, now everybody's kind of piecemealing their AI story into it. I think we're going through a big hype cycle right now. Companies are, even when you look at M&A, they're sort of proceeding with caution. I think I've been particularly interested about like what's unique IP today?

because they've seen a lot of these startups and when you sort of pull back the curtains you're like well you got some like prompt instructions here but that's like fairly easily like imagine somebody just copies a paragraph and it's like they got to all your ip like that's not that's not good um you know there's some of the providers of the data you know they know they have their own proprietary data set they've obviously get valued really well in the market

So I think there's a lot of the way companies are figuring out and valuing these AI companies. But we are going through a big hype cycle. We do see a lot of very big multipliers on the venture investments. But when you look at the acquisition side, like I said, there's a lot of skepticism still there. And you don't see a lot of these AI companies just getting snatched up left and right.

I think because of those high valuations of the VCs, like their exit expectations are going to be high.

I think you're going to see a few winners and a lot of losers in the coming years and probably a real correction around how these companies are valued. And you'll probably, like we talked about those different roles of AI engineers, you're probably going to see the classification of different types of AI technologies and that they will have their own valuation according to how strong their IP is.

Is there a certain subset of software that is getting snatched up left and right these days? Of software? Or it doesn't have to be software, but I was thinking software, some type. I think there's still a strong appetite for SaaS that's following the traditional business metrics. Because if you look at the private equity world, there's a lot of the software-focused private equity firms

But they, you know, they raised their funds even prior, like years ago. And they raised their funds on these, on this thesis that was from, you know, been like leading for the past couple of decades. That is like very much focused on the trend with the SaaS business model. And it's been very favorable for valuations.

So if you have that business that's in that 40% year-over-year growth segment and you've got good SaaS metrics, you don't have churn issues and things, I think those tend to be very highly sought-after businesses. ♪