Ex-Instacart president launches Beacon Software: an AI-first holding company for niche vertical software

Thanks so much to you. Get us up to speed. Give us on yourself and the company, please. Yeah, we we uh my co-founder Devia and I started the company a year and a half ago. What we are is a AI first holding company.

We acquire, operate, and then we grow niche vertical market software businesses that serve under underlooked at markets. We call them main street businesses. So, think about campground operations, K to2 private schools, uh construction companies, uh things that usually don't have large IT departments.

They're not the first place big tech goes to bring uh technology and innovation to. Um Divia was leaving Sequoia, I was leaving D1. We both were thinking about ways that AI was going to uh transform Main Street and that's what got us together.

And we felt like um access to the best technologies usually doesn't diffuse um evenly.

And we thought the easiest way to get the best technology in the hands of end customers on Main Street is by partnering with entrepreneurs that already run these businesses, already have these relationships and then turbocharge them with our team tech process and you know having been there and done that in scaling businesses in the past.

So um that's a bit about what we're about. Yeah. Very cool. Where where can companies that you're that you've acquired to date or plan to acquire get the most leverage out of AI so far? We we had Vlad on from from uh Robin Hood earlier.

He was talking about getting leverage at the CX in CX who's also talking about fly fishing. Well, yeah. [laughter] But getting leverage in CX, getting leverage in engineering. What What are the other kind of buckets that you think uh AI efficiency is is real today? because we've been having this debate.

You know, Chat GBT is an amazing product, but at least in our dayto-day and I think many of our of our peers, it's not exactly like replacing a full-time hire at this point. Yeah. Yeah. I uh I think Vlad definitely can use some of our products to find a spot to to fly fish.

Uh we own three businesses and campgrounds, but but other than that Yeah. Yeah. Truly. Um Okay.

So, so, so yeah, like the per perfect example because I'm uh [clears throat] I I feel like some of these categories just looking at the business and saying how do we build soft, you know, software, not even AI to make them more efficient, there's already efficiency that you could create, but uh but yeah, where where's where where where are they really getting an edge?

Let me give you that one. So, this um let me give you the campground example since we're talking about it. We uh partnered with an entrepreneur who was building a business out in Saskatoon in Canada that that served uh private campgrounds. There's 18,000 of them across North America.

Um he built great software to allow campgrounds to take reservations and you know manage bookings, manage space allocation, all of that stuff. Um we bought the business. We Mike uh runs it for us. We tend to work with our entrepreneurs and they continue to stay with the business after we we acquire it.

And the first thing we worked on with him was how can we apply AI to help him find new campgrounds. And this sounds like a very basic problem, but most of the businesses we buy do no outbound at the point that we buy them. Uh so they all of their sales happen through word of mouth and referrals.

they don't they tend not to have you know what we know in Silicon Valley the the formal go to market motions. So, we built some in-house tech to be able to first find Mike a list of 18,000 campgrounds.

Uh, then score them, do some lead lead scoring, um, some multimodal communication like when do you text them, when do you call them, when do you email them, when do you LinkedIn them, many of them are not on LinkedIn and within the first year of running that business, it's going to grow uh, more than 80%.

So, that's that's an example of of AI on the go to market side. The second thing we did with Mike is we we were doing a product roadmap review and one of the features he had built was um a lot of his campgrounds actually had boat slips and marinas attached to them.

So he built a a feature so that you could actually you know kind of reserve a spot in a marina within a campground operation. Uh and I knew nothing about the space but we started uh helping Mike. We put someone on the project with him to scope out and do calls with marina operators.

So we found him a bunch of marina operators, figured out their tech needs and then we gave him um all of our companies get cognition instances. So we gave him and his development team uh cognition our forward deploy engineers worked with him and then we launched a brand new product uh called let'sboat. com.

So like a brand new product, we built the UI, Cognition and our developers uh built the product and now we're live with that which is a you know kind of a net new company uh within within our campground operations. Um and we you know when I was building Instacart that whole process probably would have taken us a year.

We did it in like a month and a half. And I I think you know AI for code development is just table stakes now. So, I don't think it's a competitive advantage, but the matching like customer needs, customers who are in Main Street that no one's talking to with tech, that's where I think Beacons magic comes in.

Yeah, it's interesting. The at least the edge today seems to be there's a bunch of businesses that you can now build software for that previously would have not been able to justify engineering hours that suddenly can. I mean, we we experienced this.

Uh Tyler on our team builds a bunch of internal software that we use to run uh the show more efficiently and you know this was I don't know 10 years ago we probably wouldn't have uh wouldn't have been able to justify it but he can just be so productive because of that. Uh what about some of the other categories?

I'm curious with uh uh we've talked about uh alpha school on the education side using AI in the curriculum. uh yeah what what are the ways in which you can get uh efficiency in in education at kind of like a business level or are you thinking more at the actual curriculum level?

It's probably more today at the business level. So we partnered we have we we have a number of uh uh software companies in the education space. I'll I'll bring up one because I was talking to their customers earlier today. We work with a company called College Kickstart.

Uh it's based in California and their customers are K to2 private schools.

uh specifically the college counselors in K to2 private schools and they um vacuum up a bunch of data and signals publicly available to be able to help college counselors in private schools work with parents high anxiety parents and students through the matching process of hey where should my kid apply where should they apply early decision uh how should they help uh present themselves to the you know kind of target schools etc And really at the heart of it, that's a data problem.

Uh, and it's a computation problem. So, we partnered with College Kickstarter. It's one of our newest businesses. It's a it's a situation where the general manager was retiring. So, he's he's going to be transitioning out over the period of this year.

We brought in a great general manager from Whatnot and Uber and uh, his name's Tom. He was previously at Teach for America. Tom's going to work with the the old founder and and the the CEO to be able to build net new algorithms to help college counselors make better decisions for their students.

And that's like a I I have a 12-year-old daughter. The anxiety that goes around, you know, where your kids going to end up in in college is like a really big deal even for a 12-year-old, right? So, um something like that I think AI can add a ton of value. That makes sense.

uh how how much uh is traditional private equity pitching AI to their LPs uh to your knowledge? I imagine, you know, I think a ton. Yeah, honestly, I think a ton. And look, I I think we Divia and I very purposely designed Beacon as a holding company. So, we buy our businesses not to sell them.

We want to hold them forever. And there's there's a couple reasons for that. First, there's like a financial benefit to duration and reinvesting the cash flows that come from these businesses either within those businesses or in other businesses that we own.

But the real benefit is as we thought about the AI transformation, I don't think anyone really knows how long all this stuff's going to take. People have great ideas. I was listening, you had Brett on earlier, he has great ideas.

There's tons of really smart people with great ideas, but my best guess is no one can handicap how long it's going to take for the like full transformation to happen. So, we want to have patience and we want to have duration to be able to keep improving our products and keep building net new stuff for our customers.

I don't want to be held to having to sell these companies in 3 years or 5 years or seven years if that's like the moment that the magic's really going to start to kick in. Private equity can't make that promise to to founders.

So what they end up doing is, you know, and not that there's anything wrong with this, their model tends to be buy a business already with a plan of how to exit it with an underwrite that's based on a bunch of cost takeout.

Um, which they do in the first 18 months, then they start dressing it up for sale in the second 18 18 months. And increasingly smart founders, the ones we work with, they just that's just not the model for them. Yeah, that's a that's a huge advantage for you guys. Well, uh, super exciting, massive milestone.

Uh, and you're officially our our AI rollup correspondent now. So, expect expect to call as as more of this stuff is uh is in the news. Yeah. Yeah. What an amazing story. Thanks so much for coming on and break it down. Thank you guys. Take care. Great to meet you. Great. We'll talk to you soon. Cheers.

Uh, let me tell you about adquick. com. Out ofome advertising made easy and measurable. Say goodbye to the headaches of out of home advertising. Only adqu combines technology, out of home expertise, and data to enable efficient, seamless ad buying across the globe. Elon Musk's pay package has been approved.

Tesla shareholders approved Elon Musk's $1 trillion pay package. For reference, uh it it received 75% of the vote. Uh the $1 trillion stock award is