The Transaction

TT - 063 - David Boskovic
===

David Boskovic: [00:00:00] As founders, you do really try to optimize your ability to be prescient,

it is getting. incredibly hard to be optimistic enough about what's possible,

We have built an AI first Go-To-Market engine at Flatfile that can run 10 x the capacity. Every single revenue team, big and small right now . If you want to stay ahead of the curve, you need an AI engineer in your rev ops org.

So dream your wildest dreams and see if you can make them happen. And if not, put them on the shelf, not for a year, but for a few weeks. And maybe you'll see something that comes out that'll enable that.

It is changing very fast, and if you want to early adopt it, you have to take using the things at the edge of possibility pretty seriously.

Find the things that make you angry because they're not good enough. And go turn that into your marketable skill.

Craig Rosenberg: That was a mind blowing rant.

All right, so you sound great. So we're good.

David Boskovic: Okay, perfect.

Craig Rosenberg: hold on. Video and audio previews are available in the recording Sidebar.

Matt Amundson: Whoa.

Sam Guertin: See, this is why I didn't want to give you access 'cause there's just so [00:01:00] many more things to distract you.

Craig Rosenberg: Oh, whatever. Jesus.

Matt Amundson: of what makes Craig great is his curiosity.

Craig Rosenberg: Yeah. This is

Matt Amundson: lifetime learner.

​

Introducing David Boskovic, The CEO & Founder of Flatfile
---

Craig Rosenberg: we will, just kick off David. I, I do wanna mention something as we go, which is, um, you. Are a huge game changer for both Matt and I. So like, um, you know, we've, we've talked, we've been recorded on our midlife crisis, um, that we went through.

As we started to figure out that when we thought the B2B Go-To-Market playbook change, we just didn't realize [00:02:00] how drastic it was. And then, you know, Rory comes in, my office says, you gotta go talk to Flatfile. I'm like, all right. And go over there. And I'm like, holy shit, this guy. I mean, so this is gonna be, this is, we've been looking forward to the show.

It's just hard to get booking everyone. I just, I, you know, this guy's the guy. Alright, so, um, I'm gonna talk to you a little bit about the show. Here's how we do it. We got two questions. Question one, don't start until I go through everything for everyone is tell us an amazing Go-To-Market or founding story of something that you've done, you know, heroic and, and, um, amazing.

And then second we'll talk about the, you know, one to three things that you've done that are working today. So the first one is the story.

The Crazy, Raccoon-Filled Story Behind David Starting His First Company
---

Craig Rosenberg: So this is where we get the TikTok fired up. Like tell us, tell us an amazing story about, it could be Go-To-Market, it could be part of your journey founding this thing.[00:03:00]

Just, uh, just, just compel us. Let's do it.

David Boskovic: Yeah, No pressure.

you, asked me about this earlier and I was like, oh, man, what's a good story? And I, I'll say, I think the, the, the, the one that's the most fun here. actually the first company I started. Um, so back when I was 19 years old and quite a few years back, uh, not as many years back as it would be for you Craig, but, um, uh.

Craig Rosenberg: Oh man, this guy. Oh God. TikTok, here we come. All right. We are good on the TikTok content today. That was excellent.

Thank God

man. uh, but, um, no. So, uh, so, uh, I was 19. Um, I knew how to code. I, I, I won't go all the way back to, uh, you know, I started coding in high school and I built the, built some things for folks, but I was hanging out with some friends. this kid also, 19, he was telling me about, uh, he's telling me a story.

David Boskovic: His dad was like, worked for a large international bank. And [00:04:00] his dad was, you know, complaining. He was like bitching and moaning. He was like, oh man, we just got this new contract to this company in Spain. They, they were there three years behind on building this software. just told us it's gonna take us another three years and they're going to charge us another $6 million.

I'm like, you know, on my third beer at 19 years old, um, and I'm, I'm just like, me half a million dollars. I'll do it in three months. No way. Uh, next day his dad calls me. Um, and he is like, so I heard you could build this thing for half a million dollars in three months. No way. Yeah, I'm, I'm sure, I'm sure I could. Um, I, I have, I have nobody, nobody told me I couldn't, so, you know, maybe, um, so I do a little bit of diligence on it. It's actually a very reasonable project, right? It's like, this is back in 2000 and. Late two thousands. Um, and it's like they're trying to build like an internal CRM pipeline tool that looks at banking data, looks at, you know, bank usage, predicts which products people [00:05:00] are gonna buy.

Like, it's a very tractable problem, you know, it's not a world yet where like there's software for everything. Yeah. like a massive thing. So I'm like, yeah, I can do this. So, um, so I, I signed the contract. They're like, we will throw away half a million dollars at the upshot that you possibly could do this because we're so frustrated with this like thing as it stands.

I call up a dozen of my buddies, all like late teens, early twenties, and you know, I'm like, Hey guys, I, I like did it like Mafia Boss style. I'm like, I've got $500,000. We're gonna split it equally, you know, and we have three months to go build this. Are you guys in? Um, and so I got a dozen folks to go and we lived in this small little house in Birmingham, Alabama, where I had a friend.

Who was old enough to buy a house in my, we bought a house. We didn't rent it. I was like, we're gonna buy this house. We're gonna fix it up while we build this app and sell it at a profit so that it doesn't cost us anything. Nobody's losing anything on the rent here. We'll cover our food. We're all walking away [00:06:00] with equal money.

Um, so, uh, we bought this house. There's a hole in the floor that our raccoon would come out and like we, we named him, we gave him food. We had, we had, we had a ton of fun in there,

Craig Rosenberg: we named it.

David Boskovic: had three bedrooms in this house. We put bunk beds, like up the walls in each room, and we all, we were all young enough to, like, this was the craziest thing we ever did.

We, we, we, we did it, we shipped it. They, they still use it in production today,

Craig Rosenberg: No way.

David Boskovic: We pulled

Craig Rosenberg: No.

David Boskovic: app, and so I, I'd like to say ever since then, I've just been Dr. Drinking and accidentally selling software. Uh,

Craig Rosenberg: That is Okay, let's just, let's stop the show. Stop the show

that,

that is unbelievable. And they still use it today. Oh my

David Boskovic: Uh, we actually went on and we ba went, went and built like internal banking software for a couple other banks and turned it into a thing. It's actually interesting. Probably the next 10 years of my career after that were made by this one reference. 'cause this, like this executive in charge of [00:07:00] this problem, I could put like, Hey, you know, made your international bank like you need a, you need an enterprise reference.

You need someone to like back you up. This guy put him on a call with anybody. Be like, he worshiped the ground. I walked on, he was like, David, he'll do, like if he says he'll do anything, he can do it. And that just was a massive unlock for me. Every time I was like up against a big enterprise sale and I was then, you know, you know, I'm a, I'm a young guy that like nobody.

Sure, like where I came from and who to trust. I put this guy up for a reference and bam, we'd like it. It opened up so many doors for me. Uh, so yeah, I think for anybody who's listening, it's like early in their career, don't, let anybody tell you that you can't do something. Um, and, and, and sell software.

Don't while drunk. Maybe that's the,

Craig Rosenberg: Well that, by the way, today, Matt was doing a presentation for us and he, he said, what'd you say? So I inha iib some chemicals.

Matt Amundson: Oh yeah.

Craig Rosenberg: And then worked on AI for the next six hours straight. I'm like, all right. Well, [00:08:00] I think that's also a best practice. All right. That was perfect. So, um, that was, Sam ranked that story in all time transaction lore.

Sam Guertin: Uh, that's top three for sure.

Craig Rosenberg: It, it had a, it had,

David Boskovic: makes it a top two at

Sam Guertin: The, the raccoon. This is, this is,

Matt Amundson: uh, this is, this is better than maggots on a plane. I'm sorry. It's better than maggots on a plane 'cause,

Craig Rosenberg: yeah, because it had a great end.

Matt Amundson: Yes, yes. It's not

just that, that, we got a photo, you know, it's like there was a raccoon, right? I could see an animated film of this,

like this. Man,

this is good stuff.

Craig Rosenberg: Oh

man, I love

David Boskovic: left the hole in the floor to last to fix because you're, we actually liked the raccoon, so the last thing we fixed in the house, um, 'cause we, we left the

Matt Amundson: I.

David Boskovic: it to come up and we had like, like cat food for it and stuff. Yeah.

Craig Rosenberg: Oh my God. That is, uh, that, that, that, yeah, I would say. If I think about it, it had a, [00:09:00] a insurmountable challenge. Um, it had a raccoon and then it had, it seemed like it was gonna end in some disaster, and then you flipped it around. Um,

so thinking of the

David Boskovic: I mean, that's every day. every day. of my life. Yeah. Yeah. That well, that's right. You learned how to deal with the hours that you spent as a founder.

Craig Rosenberg: Yeah.

David Boskovic: day. So it's like, this is gonna be a disaster. Well, you know, maybe not. Maybe we'll pull it off. Yeah.

Craig Rosenberg: I, I love it. I love it. All right, so that was, um, well done. All right, so here's the big part of the show here is, um, we're gonna ask you to give us like, you know, up to three things that you guys are doing. Um, that, uh, are working today in this environment particularly, we like, you know, things that are surprising well with you.

Everything's surprising, so we're not worried about that. But, um, we, we wanna hear from you what's working and be, before you go, I just want to [00:10:00] say, so, you know, like, uh, guys, Scott Alber, he's been on the show seven times and he, you know, as you know, he. He, uh, interviewed you and was like, hold, he, that's when I got the text.

And your new nickname, he wrote Bosco the Beast. Um, because he was just, yeah, he was like, I'm what's happening to me right now? And he came up with this term that we, we really like, he and he presented and which is the Go-To-Market founder, which is, you know, a founder who owns Go-To-Market, like they own the product and, um, you are.

The epitome of that, because I'm pretty sure he named it after the meeting with you. So that's the, I wanna set the tone. You, you know, uh, the reason David's on the show is he's blown our minds and, um, you know, our co-host Scott's Minds, he blew our, uh, our, you know, uh, summit attendees away. Uh, you know, some, I think he had, he did a couple one-on-ones after that with like, [00:11:00] uh, heavy hitter CEOs whose CMO heard him talk.

It was like he needs to talk to my CEO. Um, so you're the perfect guest to take on this topic. What we're gonna do is hand it to you and then we will we'll, uh, ask questions and make comments as we go. But we probably don't need to because I think for the next 43 minutes, it's gonna be all you.

David Boskovic: Okay. Well, Good luck. Um, fortunately, I love hearing myself talk, so you know, this will be great.

Craig Rosenberg: Yeah. You're also another reason you're a good fit for the show. Yeah,

exactly. it's like what? Uh, no. Well, may, may, maybe setting a little context. Uh, so we're, we're series B stage, um, SaaS company raised a hundred million dollars. Um, and we had the, uh, sort of challenge and opportunity. Of doing a pretty significant Go-To-Market Mm-hmm. Mm-hmm. We started off serving SMBs and a very relatively small market, made a very significant Go-To-Market [00:12:00] motion shift.

David Boskovic: Not too much of a product pivot, but more like figuring out. We finally found product market fit in the enterprise and realized, hey, this is like a whole new world. Basically nothing we built before, um, sort of applies. And so this was just at the beginning of like, you know, chat GPTs blowing up. We started to see some of the opportunity of AI Go-To-Market.

And I'm like, okay, you, well, I with the team. I said, Hey, this time around we are gonna go AI first. We like didn't really know what that meant, but like, if we can use ai, if we can't apply, um, this to our emotion, we will. We're not gonna just go and sort of do things the old way. Um, and for us, you know, we're a relatively small team running an enterprise sales process.

You know, enterprise sales is generally like very heavy handed and

Matt Amundson: Yeah. Yeah.

David Boskovic: Um, there's so much for customers expect from you, sort of every step of the way from building POCs that like are basically a whole fucking implementation before anybody signs a check. Um, all the way through, you know, 47 stakeholders all with [00:13:00] their own weird, you know, problems in the process that have to be.

Uh, worked through. so, you know, but then you start, you know, thinking about sort of ai, you're like, oh, maybe, maybe we can, maybe we can solve some of these. Um, so that, that's, that's the frame, uh, what we've done. Is, I think we've delivered for ourselves on that, which is we have built an AI first Go-To-Market engine at Flatfile that can run I would say, you know, compared to the, the pipeline I've run in previous motions at Flatfile and, and out of the comp, uh, other companies we're running about 10 x the capacity, right? So pipeline to sort of employee ratio, extremely low. Um, and this is across the entire motion, right? So marketing all the way through sales.

Um, on, on marketing, we've, you know, we're running an ABM motion, highly verticalized, highly focused on the target accounts that we have, and we're doing something that would generally be pretty expensive, but we're doing it with, you know, deep research and we're doing it [00:14:00] with AI almost entirely. Um, so, you know, I I, some of you I may have, yeah, I don't know if anybody's listening has seen, seen any of the stuff we've built here, but, um, you know, I'll, I'll maybe describe it.

Um, so starting from the beginning of the funnel, we have the advantage of knowing every future customer of our product, right? So that's, that's a great setup of an ABM strategy, right? If you have a relatively finite market, you can kind of pick your future buyers, you know who they are. Then the, then the real problem sort of in Go-To-Market becomes the efficiency it takes to get in front of those people and make them want to buy your product, right?

That's like pretty much it. Um, now not every company has that advantage, right? Sometimes you have a much broader sort of. Like market, it's a lot harder to like actually pick out your future customers, but we don't have that problem, right? So if you're a company that sort of like knows how to pick your customers out of a list, then you know, ABM or now ABM is interesting to run, usually best run and sort of a verticalized motion

Matt Amundson: Mm-hmm.

David Boskovic: you know, it, it, it, [00:15:00] know, more heavy handed with like cus company research, AEs, SDRs, et cetera.

We asked this question so. Immediately after deep research launched the Chat's Deep research, the moment I got access, I was like, I want to see Deep Research can go and figure everything out about a vertical Go-To-Market. Play around, you know, a finite number of companies, let's say a hundred or so.

That is exactly the right sell for this customer Now for us, while. We have, we know all our future customers. They, we have a problem where each customer has a slightly different applied use

Matt Amundson: Sure.

David Boskovic: different value proposition. It's not all the same or CRM, it's like, Hey, I'm selling you A CRM.

This is gonna make your salespeople better. They're gonna do their job. You're gonna close sales faster. Same sell for pretty much everybody. For us, we sell. know, a data preparation product, right? And data preparation is a different problem in each industry, right? If you're selling into insurance, well, data preparation is getting data from [00:16:00] brokers and going through insurance data and dealing with thousands of PDFs and blah, blah, blah.

you're selling to a, you know, enterprise. HR software company, right? It's all about like pay stub history. It's all about 80 different sort of like tax rules that need to be, you know, worked through. So like the cell, right? Everybody comes to the door and they're like, Hey, but can you do this? That's the thing that we care about, right?

And if we open the door with that, like, Hey, we understand your industry. We understand this key specific thing about your industry, the problem that you have, and here's the like thing that you need to know we can position the product, the demo. The, the, pitch, the demo and the entire sort of onboarding experience towards that cus customer or that vertical, then we have a brilliant process that the customer kind of feels like we understand them every step of the way.

But that's hard, right? A hundred verticals in, that's hard to maintain. Right? Hey, you know, se we have a hundred different, you know, verticalized demos to run and they're all really deep and they all really understand this problem. That's like an [00:17:00] unreasonable thing to do with three se. Right. Except we do it right.

Hey, you know, let's, go and, um, let's go and do customized outbound to 10,000 companies each where each outbound actually is not, it's not like this stupid AI like, oh, I saw you like posted on LinkedIn yesterday. It's actually thoughtful. It's actually right, and it actually understands the problem.

So it's back to my first exposure to sort of ChatGPT deep research. I was like, let's see. If you can take a company and take a market and go and develop a thesis on Flatfile for that company that I can send raw to the CEO of that company. And it doesn't seem like bullshit, So I had ChatGPT BT start on a market.

I, uh, I did, uh, field services, you know, so your world of service, highend job or house Call Pro, that world we had, we had a bunch of customers there. And I was like, Hey, you know, go research this market. Um, and I gave it a prompt, like, you know, our problem in the market, what we solve for, you know, what I was looking for, et cetera, wrote a 40 page report on the market and I sent it to one of the [00:18:00] CEOs of the industry.

I didn't tell him where it came from, and I just said, Hey, you know, I got, I found this report. I'd love for you to take a look through it. I'm really curious if it's like, if you find any inaccuracies or anything we're thinking of like building something around this. a note back and he was like, dude, I learned so much from this.

Where did you get this? And I like, that was, for me, it was like it passed the sniff test, right? Like when you're trying to do like an automated strategy, usually you're worried about the quality of what comes out of

Matt Amundson: Sure.

David Boskovic: And so then I was like, this has legs. So we built a whole strategy around the fact that AI had arrived at a point where it could.

Couldn't just go and sort of do orchestration for you, but it could do real valuable research that you otherwise would've had to hire an MBA for spend them a month building this report, which is still reasonable economics, frankly,

Matt Amundson: Yeah. Yeah.

David Boskovic: you, most people would spend 20, $30,000 on a report like this in order to go after a hundred million dollars industry, right?

That's a very reasonable thing to do, but yet ChatGPT can do this. And then I was like, okay, well okay, if I can do this. [00:19:00] This easily on the market. What about the companies? And so we got to a place where we, we ended up spending tens of thousands of dollars on, on perplexity credits to go and research every single account and incorporate every single data point that we had already learned about that account and research everything publicly and develop a sales thesis on that specific company.

And the ROI propositions rank them in terms of likely pain to the customer and Bill basically, Hey, this is the game plan for going into this account. and it's all tractable, right? Are there tools that exist for this? not really.

Until, actually there's one that I think just launched yesterday. It's like called end game, but like literally like these tools are just emerging in market and I told the team like you're, we want to move before there's a tool available.

We'll go do stuff that we build internally. We have AI engineers that build our product. We're gonna use those people to go build the internal tool. That does this and eventually six months beyond. You know, once we get this out, someone will launch it. We'll be able to stop maintaining it and use their product.

But we're gonna [00:20:00] do it first the hard way. And it turns out you can vibe code these tools like over a weekend. Like they're not that hard to build. If you have an AI engineer and you have a clear idea of what you want to do operationally. They're reasonably easy to build. So, yeah. I'll, I'll, I'll pause there 'cause we're Yeah, yeah.

It's like the, like can you, the, the fundamental question is can AI do it? And these are easy to validate, right? To pick any problem you have in your business and Go-To-Market. Can AI do it? The next question becomes operational. Can we actually do it at scale? And if you can solve that, right? If you have, you know, an AI engineer or you have

somebody who can hack out a process internally, you can actually very quickly start operationalizing AI in a way that gives you unreasonable advantage in market right now. Right? Like the motion that we're running would take easily a hundred

Matt Amundson: Yeah.

David Boskovic: run at the level of diligence and scale that we are.

Matt Amundson: So I have a, I have a bunch of questions. I mean, first, a, a few observations. One, most of the time when you hit PMF and you're like, Hey, I've raised this much money. You go, nice. I'm gonna build a, you know, [00:21:00] 75 person Go-To-Market team to go execute this, right? Like, that is the playbook, right? We need a content marketers, we need product marketers.

We need demand gen marketers. We need customer marketers. We need marketing operations. We need a sales leader. We need a multitude of sales teams. Are specific to each one of these verticals or potentially by size, we need managers to support those people, sales ops, to support those people, et cetera. And what you've done is sort of one, in my opinion, and, uh, I don't, I don't want, I don't wanna paint you with a brush.

If this is an accurate, like the same version of you that said, fuck it, I can build this software for 500 grand, was the same person who said, fuck it. I could build all of that with an AI tool. Right. So one that takes a ton of courage, and generally that's not a CEO's job, is to be like, Hey, I'm gonna build this, like, Go-To-Market strategy.

And like, I'm gonna, I'm gonna be kind of the, the, or I'm gonna hire some people, but like, I'm gonna be the straw that serves the drink here. So one that was courageous as [00:22:00] shit. Uh, so that's my observation. My question is, is like. Where I find research reports to be the most impactful is the more specific that you can get.

Right? Like it's not enough to be like, Hey, here's a state of marketing for all companies that do marketing, right? Like you have to break it down into some level of segmentation. And you had talked to initially about like maybe a group of a hundred companies. Like where do you find. Like how specific do you get down to before, like you potentially run the risk of building something like this that's like, Hey, this is just trying to apply to too many peoples and overall it's lost its effectiveness.

David Boskovic: Yeah. Yeah. Well, first of all, the comparison to my first startup, um, is kind of flawed since there is no raccoon. Um, so

Craig Rosenberg: Matt. Matt wasn't listening. I'm just gonna say it. Matt wasn't listening.

Matt Amundson: Oh, come on. Oh, I referenced

it.

David Boskovic: um, but, um, okay. So, [00:23:00] uh, my, my position on this is. Uh, I was, uh, talking to someone recently. I was like, I actually think you, we, we have a lot of concerns that AI is going to sort of increase the amount of, sort of like drivel we see everywhere. But anytime you make it cheaper to do something good after some amount of time, luxury gets more accessible.

Matt Amundson: Sure.

David Boskovic: Right, so, so will you, if you really think about sort of like what will happen in sales, sort of by virtue of the economics of what is happening right now, like over some amount of time, right. Things that we consider luxuries today. become more accessible to everyone. So, for example, today it's a luxury to go and do deep research on a, well, uh, maybe not today, anymore, six months ago on a, on every single target account where you know everything about their history, you know, everything about their product, you know.

Everything. You've read every page of their docs, you've developed a thesis. You would only do that for your [00:24:00] most valuable accounts. You're going after Walmart, you're doing that, right? You want to go nail that $7 million get, yes, you're doing that, but are you doing that for the $50,000 get no. It just, it costs you $50,000 to do it.

Why would you fucking do that? It makes no, no economic sense. So the economics are shifting. So then the question becomes, okay, operationally how do you actually do the thing that's a luxury, uh, and, and like structure it, right? So for me, I was looking for a unified application of the product where the message could be coherent and I could do diligence on the message, right?

So what's interesting is the best, the best, most luxury right is individual accounts, but that's too detailed. I cannot review 10,000 different accounts for quality, but I can review and audit a hundred different micro verticals each with somewhere between a hundred and a few hundred accounts that I can do.

Right. I can look at and I can do quality review on it. And one of the things that like I think we'll always have is like we'll always need [00:25:00] the. always is maybe a, a big generalization, but I feel comfortable when I can look at something and I go, that passes the bar, right? The bar that I have for, you know, this sort of play.

So we choose micro verticals of no more than 500 companies. Somewhere between 100 and 500 companies. If you find a vertical that has more than 500 companies we break it down further until we get down to a reasonable set of customers that have a reasonably homogenous sort of use case and sort of messaging.

And that's where we go and then do most of our investment. This is our messaging strategy. This is how we pitch the ROI. I'll give you an example. Like there's a sort of a group of software companies that serve like cannabis and regulated

sort

of, youu know, stuff like that. And their big problem actually is that they have to, by regulation report data to the states every night, like nightly sales reports effectively.

And every state has a different file format requirement, right? So if you're selling to those people, [00:26:00] you want to go and solve a problem that is causing the misery. Tell them that you're solving that problem. That's it. The deal's done

Matt Amundson: right.

David Boskovic: else is like, cool, well we'll get that. But like that's the one problem that you only get that if you actually know enough about the industry and do you think I am the expert there?

I'm not. But then you go and have deep research, go and do this and other you're like, oh, of course. We've seen this come up in three sales deals and look, looking at our history, we see that we've lost. Every one of these deals. I wonder why It's probably 'cause we didn't know about this and didn't know how to pitch this.

Let's adjust. Right. So, you know. Yeah. I, I, I think getting down to a small enough vertical that you can do diligence is the biggest problem. I think over time we'll get smaller and smaller. Right. But you have to figure out somehow to how to operationalize the quality control on it.

Matt Amundson: Nice and extending beyond just like the research report is. Then are you then like using that data to tailor the deck? Tailor the message, tailor an ad, tailor a landing page?

David Boskovic: Everything. Um, so we are, this is like, we are [00:27:00] incrementally doing this, but we are getting, I would say we're at a 60% deployment of Yeah. Um, fast forward to a hundred percent deployment. We will have a beautifully designed v landing page for each micro vertical with a AI generated video that's based on a core video that is undetectable, right?

It'll have product walkthroughs. The product shots though will be verticalized. Weird. How do we get a product shot that was verticalized for that particular industry? Right? The data examples, everything in there, it'll be production quality, right? Your ENT entry to this product, it'll pitch the exact scenarios, right?

And we'll, we're doing a bunch of like clips of, of, of, you know. Generalized talking points that can be incorporated into that so that it feels like a production level thing, but it was AI generated. Um, then in each of those, right, many of them, we actually have anchor customers, so five or six very strong anchor customers that we want to use.

And so if it'll go and analyze. Every conversation we've ever had with those customers. Create the use case analysis, put it on the page detail how, how customer x, y, and [00:28:00] Z uses us. And, you know, we'll do some auditing and report on that, but we have a structure for generating that content. Uh, and then it generates the like different sort of key use cases in the industry and different break functional breakdowns of how you would onboard and apply and a custom demo.

Now, custom demos, we already have nails, we have, we have, we have this AI script that takes a given logo. And auto generates a fully custom configuration. Example, data, full script. Start here, say this all the way through the demo so that for that company it hits perfectly. And I will say the conversion rate on the on, on being.

When you go into a first conversation with somebody, you have 45 minutes and you come in, you're like, I researched your company. I did some analysis on like, sort of like what you're doing in market and, you know, we might be wrong about this, but based off of what we saw, it looks like the first problems you have, the primary problems you have, that we might be all helpful with our X and Y.

So I put it, I put together a demo today for X and we had our team do a little bit of, you know, work prior to this. And, um, I wanna [00:29:00] walk you through it. It should be, you know, maybe, maybe it's not perfect, but it should be, um, you know, a pretty good example of what the product does more often than not. It's like you're in their fucking brains.

You're, they're like, wait, that's our data Yeah. I don't even, I didn't even know that. But AI went and read their docs and found their API and actually did like a perfect example of that product. And people are like. luxury, right? You do that for your biggest accounts, right?

But Yeah. it's, it's a luxury that can be afforded for everyone. So now you can come to your first demo and you can see the product basically implemented already for you. And surely you don't leave going, wow. This is gonna be like, man, I'm gonna have to talk with our team. I'm not sure if you can get the engineering resources, because what you saw on in that conversation is.

just want it tomorrow, you wanna start using it tomorrow 'cause it's close enough. And that that difference between a generalized demo and a more verticalized demo to the customer. It's everything. Um, and then being able to align the talking points in a demo where [00:30:00] you're incorporating like ROI proposition and product in the right flow.

So it's not like, oh, I'm gonna go pitch you for 15 minutes and then give, give you a 15 minute demo and we're gonna talk about next steps. It's like the smooth flow where you're like, yeah, let's get into the product. It's high energy. Let me show you this. And you're just like, slipping in these like ROI.

And by the way, this is how we solve, you know, the, the onboarding people, your customers, you know, we, we've seen this with other customers. It's not 30 days anymore. Three fucking seconds. Um, like whatever you're trying to pitch in there, right? You're incorporating the pitch. But if people, if you get to sneak it into a demo and the demo is unique to the customer, and so then it feels like an incredible people leaving, you're like, wow, that product, it's looks so valuable.

Matt Amundson: Yeah. yeah, well, Yeah. That's the show that we wanted to put on. And yes, it will be valuable. By the way, anybody listening, FFEL is incredibly valuable. We're not, but we're just good at, we're also good at selling it,

Craig Rosenberg: That's amazing. Honestly, Matt, by the way, that was a good question. I just wanna let you know before you go, David, I just thought of something really like from this show, which is, I think Sam should [00:31:00] start, um, coming to the show in a raccoon costume. You would be a raccoon because David just reset the standard here on like, what's cool.

Um, it is the, the, it's, you know what's incredible about everything you've said? Everything's undeniable

Matt Amundson: Yeah.

Craig Rosenberg: and, and you have, but there's something you didn't say, which you've told me before, which is true, which is demos can suck 90% of the time.

Matt Amundson: Yeah.

Craig Rosenberg: And they're typically very unprepared or starting from scratch.

Even folks that prepare now using like a, you know, a ChatGPT to bring up three things. Hey, our customers, they still do a terrible demo.

And you know, what you made me think of was NetSuite back in the day, they had a 50% conversion rate when they demoed. But here was the thing, [00:32:00] they, back then, you, you know, they put people on it.

They spent so much time to make the demo look like yours, uh, that it killed it. And what you're doing is akin to that at scale. And it's, it is such an important point, man. I mean, I'm just like, bam. That was, That was,

a great rant.

David Boskovic: on fundamentals. Like we are already all doing these things badly,

Sure. Yeah.

we attempt to do them. right?

So, so like the bar that the in many of like, I'm just running ABM, right? I'm not doing anything fucking like different, right? This is ABM, but at scale with a lot of attention to each account in ways that previously were unaffordable, right?

Like it's the same motion, it's just done differently. You're like, oh, we'd like to give better demos. And better demos. And better demos, of course, right? You do training, you do all this stuff, but then you're like, okay, but what if we could do something that we would think is impossible, like have an se go build the demo prior to the first [00:33:00] call before we even qualified the customer, where we might have wasted, you know, six hours of that person's time for absolutely no reason, right?

Probably are not gonna do that, but now you can, right? So again, back to this point of luxury, right? These are luxuries. Already give to ourselves for the most valuable accounts. What is changing is that you can give it to yourselves for the cheap what For lower value accounts for. Previously the economics didn't make sense.

Craig Rosenberg: Yeah.

That's me, let me ask a, a couple questions. 'cause I imagine that, like our listeners are, are wondering these things now, like the basics, like how long did this take to build? What, how are you, like, how much. Money are you deploying towards this? Where are you not spending dollars like you were potentially in the past or where you might have been had you gone out and hired this team, and where are you redeploying this or where, where are you redeploying those dollars in order to make this a reality?

David Boskovic: Yeah. Um, okay. Lots of questions in there. I, one. How long did it take? [00:34:00] Um, look, uh, I, I think the first versions of everything happened very quickly, yeah. Yeah. Um, you know, uh, my technical chief of staff and I, um, he's a previous CTO, we like hold up for a weekend, um, north of Marin up in wine country a little bit.

Um, and we are just like, we're gonna spend a weekend and we're just gonna like ship this and we call it Flatfile. CIA, we're gonna ship this Flatfile CCIA that can go research every customer. So, you know, it's a little bit of engineering work. Um, it, it, it's a lot of prompt engineering like, Hey, how do we get the right thing?

And then it's like. $10,000 of Perplexity credits. Uh, like just Go Ham, right?

Matt Amundson: Yeah.

David Boskovic: let you, let, let you do it. Um, and then you're like, holy shit, we have this. Um, then we did something that I think I haven't seen yet, and I'm like, I'm waiting for someone to build someone to build this.

So anybody out here who's building a should Go-To-Market software company should do this, but The memory layer on top of your collected wisdom.

Matt Amundson: Mm-hmm. Yeah.

David Boskovic: is you, we've had tens of thousands of calls with prospects [00:35:00] over the years. A small percentage of those right have already become Flatfile customers.

Right. The rest should still become customers. And in those calls we learned things. Right? In those calls two years ago when we were talking to the head of, you know, implementations at X company, they said, yeah, you know, we do 500 implementations a year. We charge our customers $32,000 in implementation.

We have a team of 42. Uh, this is a nightmare. Right? And for whatever reason, they didn't buy. Maybe you think, but that's. Memory. We know that as a business, but do we, will that ever be truly known in existing CRMs? That kind of insight disappears very

Matt Amundson: Yeah. Yeah.

David Boskovic: but it's an important piece of knowledge about the account.

So what we did is we have. Uh, we have an AI sort of crawler that goes and looks at every gong transcript, every email, every every CRM note and it, it is trained to find the salient points of information about each account. So we actually have that available. When we are doing new sales. [00:36:00] So when you go to a new sales call, it actually now already remembers from the call two years ago certain qualification information that it can then, you know, give you as a new sales person.

So you're coming in, you're like, yeah, we talked to your team a couple years ago. We heard that, you know, there's a huge problem. Is it still the same? Has it grown? Is it less? And you're like. that right. The cost of getting that would be very high,

Matt Amundson: Yeah.

David Boskovic: the sales rep took that exact note, it was like top of funnel, like top of the account.

You have to go watch the gong call. You try to get that information. So teaching an AI what to think about as a salient point and then adding it to a memory engine helps us not only understand what's out there market, but in a lot of cases there's a lot of information that we've collected through our work.

Right. Our work might have just been failed sales calls, but that was work, right? There are conversations, it was support tickets. Someone from that company, you know, chatted in a year and a half ago and was like, Hey, I'm playing around with this. I'm trying to use it for X or Y information memories. Now, if you cared about that particular account, you would go and do the research.

If it was a large enough account, you would go have someone read [00:37:00] every single fucking conversation you ever had in order to find these points. So again, to the point of luxuries, right? These are luxuries we would. Already do for the most valuable accounts. Now you can get them everywhere. Operationally, though, that's quite

Matt Amundson: Mm-hmm.

David Boskovic: Operationally it's a lot of API connections. A lot of just feeding data in, running API, uh, running AI analysis on it, playing with the prompts to figure out what's salient, storing it in a database where it's available for retrieval. And one of the things I realized, and I had talked about this, uh, uh, at Saster a couple, uh, uh, a few weeks ago, was.

Every team, every single revenue team, big and small right now needs to hire a single person. If you want to stay ahead of the curve, you don't wanna be subject to waiting for the SaaS companies to evolve. You need an AI engineer in your rev ops or rev ops org. You have to have someone who can go and understand how to prompt engineer, understand how to apply.

All of these tools against your rev ops, it is the highest leverage role you can hire. They need to be someone who could vibe code shit over a weekend, right?

Matt Amundson: Yeah.

David Boskovic: to be hackers. Can't be like the, the, the cerebral AI engineers, but like [00:38:00] find that kid who's like out on, out on like Twitter being like, oh, look at this thing I built with lovable.

Put that person in a rev ops or pay 'em a fuck ton of money and they will go run circles around. Everything you've ever done in rev ops in minutes, it's not hard to say, Hey, you know, ChatGPT, connect to my Salesforce and find every, whatever. Like, it'll build the integrations for you. You just need to operationalize doing

Matt Amundson: Yeah.

David Boskovic: you can imagine is actually doable

Matt Amundson: Mm-hmm. Yeah.

David Boskovic: but you don't. Most people, like I've, I was talking to this company, massive public company, and the, their head of rev ops was like, oh man, I love, I saw what you did. I love what you did. Like how do we, like, how do we get it? Like, can I license it from you?

And I'm like, you, you don't wanna pay me for what I built? You need to go hire this person. She's like, oh man. There's like, I don't, I don't know how we would ever find that, et cetera. And I realized that this is going to be the alpha for people who are early adopters on the like AI Go-To-Market. Spectrum.

If you make this higher in your org and just have someone who you can just have an idea and be like, Hey, what if we could do [00:39:00] X? And that person can go within a couple days, make it operationally happen. You can, you can do all this stuff. Like everything I'm saying is like just doable. Like it's not like you don't have to wait for a product.

You don't have to buy things. Yeah.

Matt Amundson: Yeah. And if you hire that person, you should contractually obligate them to list their title on LinkedIn as like janitor so that they don't come get poached from you.

Craig Rosenberg: yes, that's right. No, you're not allowed to have a LinkedIn.

You're not allowed to

Matt Amundson: You have to be a ghost. Yeah. You gotta be a

ghost

Craig Rosenberg: By the way, that was a mind blowing rant. If Sam doesn't cut that, he's gonna have to wear a raccoon suit like out on weekends. That was,

yeah,

that was unbelievable. Oh Lord. Um, so, um, what, uh, you know, now that we're at, we're like, what do you do?

Going forward, you know, like are, you know, how do you add, how do you think about adding people now if [00:40:00] you, even if you do right? Or are there things that you still haven't, that you've dreamt of but you haven't put in play? Like what does that look like?

David Boskovic: Yeah. Very interesting. Um, frankly, I think anybody who's, who thinks they know the answer to this question is an idiot.

Craig Rosenberg: Yeah. Matt. Matt does. But yeah.

David Boskovic: I have, I, I have, yeah, marketing people, they think they know

everything. Uh,

Matt Amundson: three months I have.

David Boskovic: um, yeah, yeah, yeah.

Matt Amundson: Realize that I know nothing in the biggest way possible, like

I am back,

back to being an amateur all over again.

David Boskovic: yeah, as founders, right, you do really try to optimize your ability to be prescient,

Matt Amundson: Mm-hmm.

David Boskovic: that, that, that, that is, that it's like part of, part of the alpha. Um, and I do think it is getting incredibly hard to be optimistic enough about what's possible, right? Because like a lot of what we do is like coded at this point, certainly in the evolution of the company [00:41:00] for, for risk avoidance, right?

Try to find a series B CEO that's gonna blow up their like revenue org and rethink everything, right? It's very hard to think about doing. and, and, and so. The problem is then we, then we actually tend to be pessimistic about, oh yeah, cool, these cool AI things are out there, but like, man, it's gonna be a while before they're, you know, operational and like, look, look, it's like full of errors.

Nobody, nobody cares. Like, like, you're almost like coded to want to think that it is not going to work. And in reality, the things that are, every single model release and the things that are becoming market available, if you can access them, are so much better than I am, like as, as a CEO of an AI company than I sometimes am able to predict at this point, I've just gotten, I've told the team, I'm gonna assume your best possible scenario and then double it until at some point the curve flattens out because right now we haven't found that yet. Things are still getting better and more possible at a pace that, so if it's not possible today, wait two weeks, four weeks, it'll probably be [00:42:00] possible.

Then assume it will be possible. So dream your wildest dreams and see if you can make them happen. And if not, put them on the shelf, not for a year, but for a few weeks. And maybe you'll see something that comes out that'll enable that. Um, it is. It is wild how fast things are moving. So I don't know where things are going.

Right. All I know is that it is changing very fast, and if you want to early adopt it, you have to take using the things at the edge of possibility pretty seriously. Uh, for me, our, our theory around hiring is an amplification sort of strategy. So if you have the three different ways AI's being deployed, right?

You have AI that tries to replace people, oh, you're an AI SDR, right? that's interesting 'cause it's anthropomorphization of AI into a previously human role and AI so far has not been able to demonstrate that it can do all the things that a human does well.

Right. It can do some things well, but there's certain things that humans do that are particularly human and are important in the process.

Right. So I think anthropomorphizing AI is an easy way to get that [00:43:00] feeling of, oh, it's not quite there yet, right? Like, 'cause it's not there, but there is an artificial bar that you set

Matt Amundson: right

David Boskovic: there. You're like, well, I'm waiting for AI to be a good AI SDR. Well, okay, you'll be waiting longer, probably not forever, but you'll be waiting longer than having AI that can go and do great deep research on accounts that maybe previously was a job of an SDR, so we. deconstructed every role at Flatfile. This is across the entire organization because when you look at traditional human roles, right, they're optimized around a peak of skill. And then what you do is you like amass as much responsibility around that skill before it reaches diminishing returns, right?

So that that's how we design human responsibility, right? AI doesn't have that problem, right? AI has thousands of individual points of expertise, right? There's no peak, right? So it's all, it has lots of individual points of expertise you can apply to it, and your humans, right? Can focus more on their peaks of expertise, right?

So taking away the responsibilities of diminishing returns from every role, handing those [00:44:00] to AI to do and allowing humans to operate AI, turns humans into amplifiers. And I think this is one of, oh, I said there's three ways, right? So we take the amplification approach.

The other is the augmentation approach. This is what you see in first adopters of like engineers they're like, oh, I'm using like Claude code. And I'm like doing same thing I was doing yesterday. I'm just doing it faster and better, right? So, you know, your, your content marketers ship click bait faster, right?

Your, your engineers ship bugs faster, right? They're doing all the same things they were doing before, just faster and at greater scale, right? So I think that I, I think that is generally the first way we step into things, I do think it's shortsighted and it leads to like a volume of bad outcomes. I think in order to step back and build an organization that's designed to take advantage of this future, every human role, every responsibility, every measure of performance, and every judgment of expertise has to change because otherwise you're hiring humans to do jobs that they will not have in six to 12 months.

Right. That's stupid. That's shortsighted and you're either an asshole or you're you're blind, right? So, so you [00:45:00] should be hiring people to do the things that are, that do the best of our ability to foresee the future continue to be strengths that humans have, and make sure you're hiring people that are able to take advantage of AI doing things.

Now, the problem in all of this. it's like, it's like managing a team of a thousand people, which is actually operationally very hard. This is where organizations go to die, right? You have to orchestrate so many different things in the business, and the orchestration of those things can be where it all falls apart.

So it's actually not that easy to take advantage of this. So we found that in every area of the org where we go and build orchestration layers, right? We think of it like the AI requires you to invest in the operational layer of your business, treat your company like a software product, and debug it like that, right?

Like there are functions that need to run and so now I need a tool that goes and does this, right? We do this in onboarding, we built this internal tool that tracks our onboarding, writes out the customer spec, it analyzes every line of code that's shipped during the implementation to see whether it's violating the spec in the original onboarding, all that right could be done the hard way, but if you [00:46:00] start operationalizing it, you get to start focusing your team on the most valuable points. So what we are hiring for are two types of roles, right? A executives who can think about the world operationally, right? So I actually think the future of a lot of revenue orgs like lives with rev ops leaders or rev ops adjacent people.

People who think about the machine that is the revenue machine, right?

More than, Hey, you know, I can go and build an organization with a VP here and a VP here and a direct three directors under here and they're all gonna go do stuff. What is the machine that produces good results? Um, and that I think. Is, is is given to people who think are systems thinkers and has to be, 'cause there are so many new sort of operators in the org that have to be coordinated in order for that kind of amplification to happen.

So I think of it as an operational problem. So executives at Flatfile have to be people who are systems thinkers and can design the system that AI and their teams can work together in. And so I'm actually hiring for a revenue executive that fits this right now. I want someone to take this sort of like baton from me.

You know, I [00:47:00] think we have the first version of this that is, is working well, but I there's so much further it can go, I want someone who thinks that way and can sort of build that. So this is an open, open ask for anybody who's interested to please DM me on LinkedIn or Twitter and let's chat. Then for team players, right, you find the people with taste This is the one thing that I found that like from an organizational perspective that you have to hire for now. because when you thinkbout what's left, at the end of the day, if execution is a commodity, then what's left is judgment.

Matt Amundson: Hmm.

David Boskovic: and the judgment that you trust. So what you hire for are people with good judgment that you trust to make decisions, because execution is no longer the bottleneck.

You're not hiring a bunch of people to go and execute on 400,000 pieces of content anymore. You're hiring someone who has good taste. You're like, I know that the content that that person underwrites is going to be good.

I know it's going to be high quality. I know it's not gonna sound like bullshit. And that's the, that's what you hire for.

So what I [00:48:00] found in every IC role, I start with looking at the judgment of that person. 'cause execution at this point is different, right? I don't have to, I don't have to be like, oh, how many content pieces can you write a week? That's not, that's not the commodity anymore. So that we hired a head of content, right?

Eventually, 40 people we interviewed for this one. This is a great, great example of this. Content marketing is one of these roles that historically. Frequently, you just, just turn it into a content monkey,

Matt Amundson: Yeah. Yeah.

David Boskovic: you be, be, if you're a good enough content monkey, eventually you get to manage the other content monkeys.

Uh, but some, some people, right, like they, they're, they care about it. They're like, no, that's like, I have a quality bar. And this person came to the interview, they're like, here's my, this is what I give ChatGPT to make sure ChatGPT produces content that I love. And it was good, and the judgment was good.

And so when we dug into that, we're like, okay, this person understands the new problem in the world. I have judgment, I have taste, I will hold a bar and I'll figure out how to operationally execute on that. But every, every IC needs to, like, after one of my presentations on this.

I was talking to someone, they were like, [00:49:00] how do I make sure that I'm employable in the future? I'm like, find the things that make you angry because they're not good enough. And go turn that into your marketable skill. Like if you care about something, if you're like that, just that, that's not what, executing that will be the easy thing in the future and people will hire you because you have that opinion. And the people who have no opinions and just are, are, you know, roughly like energy conversion machines into like, like, uh, practical sort of utility. Uh, like that's not that the I'm very bearish on the, on the value that will have to the world.

Matt Amundson: Yeah.

Craig Rosenberg: Okay. That what? First of all, by the way, Matt, I'm just gonna compliment you. You have taste, you do, you have good taste. And the taste thing is so amazing. I, Sam, if you don't cut that, you have no taste and you're wearing an a raccoon outfit seven days a week.

Matt Amundson: Yeah.

Craig Rosenberg: That's amazing, man.

Sam Guertin: Fair enough.

Matt Amundson: you have to wear the tail.

Craig Rosenberg: Yeah, that's true.

All right, well that [00:50:00] was the most, uh, uh, the least I've talked. The most I've learned on the transaction.

Matt Amundson: Yeah. If this, if this isn't our most listened to, uh, episode, I'm gonna be in shock.

Craig Rosenberg: You'll wear a raccoon

Matt Amundson: I will wear a re.

Craig Rosenberg: raccoon mask.

You

Matt Amundson: but why not? Maybe a Davey Crockett hat like, you know, I'm gonna Disney this week.

Craig Rosenberg: that's a good in-between, by the way, quick side observation, so David likes to go to uh, uh, places to code and to figure this stuff out. He was in Birmingham, Alabama with a raccoon and then north of Marin.

That, those, those were two fun

facts.

David Boskovic: Um. do actually, do a lot of destination trips. Like, uh, we're a remote company, so we like, want to get together we have like, we're like, get an Airbnb somewhere that's like reasonably sick and just like go, like, hang out for a couple days and get shit done.

Craig Rosenberg: Yeah. That is amazing. Well, look, man, that was exactly, actually, that was beyond what we had hoped for, um, [00:51:00] like in a, in a big way, like the biggest way. So, um, I mean, I don't even know what to say. That was, that was so good. And so I would, I just wanna say, well, actually I can, I know one thing to say. Thank you.

Matt Amundson: Yeah. Thank you. That was

Craig Rosenberg: And, um, yeah, that I could, I needed, I should, I looked that up on ai. It's like, just say thank you. Um, and then, um, and oh, look at it. Oh my God.

David Boskovic: you're just a, a primitive energy conversion machine. Ma machine. Okay. Yeah.

Craig Rosenberg: Well, I certainly can figure out the things that make me angry, uh, um, but, uh, Sam, excellent work in the back. Is that supposed to be me?

Sam Guertin: Uh, it was originally you and then I, I fed it, a couple images of David and said, turn it into this guy.

Craig Rosenberg: It's if we had a baby. Oh

my God.

Wow.

Dude. Jesus.

Matt Amundson: Your baby loves raccoons.

Sam Guertin: I, I, this entire time I've just been thinking of [00:52:00] the, uh, the Gopher and Caddy so, so

Craig Rosenberg: it's a

David. David

David Boskovic: a crave it. Crave it in the raccoon.

Craig Rosenberg: in the Carl the groundskeeper. Uh, uh, body type. Wow. That is unbelievable. All right, well, that's the perfect way to end the transaction. Thank you, man. That was, that was perfect. I mean,

honestly,

David Boskovic: good stuff. [00:53:00]