Episode Transcript
Hello. It's Andrew from XO capital, where we buy small software companies. And this is kind of our weekly update of what we're up to. And this week we're a little busy. We bought our This is our eighth acquisition. It's called support guy. And it's kind of our entry point into this this sort of G P T mania that we haven't been avoiding, but we haven't found a really good target for us to acquire in this space. And this was a a relatively small bet. And we'll get into why we did that in a second. But interestingly, this is the same dude that built cold DM.
Which we bought a few months ago and has been a great investment for us. And we're really excited to acquire this from the same person. This was the first time we've ever bought something from somebody twice. I mean, we've only ever done. Eight of these. So it's not surprising we haven't done this before, but it made real diligence really easy. I mean, the whole process was, was really quick. Did most of it over Twitter, DMS and didn't have to do as much diligence before there was a lot of trust.
And, you know, we already had the guy's payment info on file and, and transferred everything. And he kind of knew the process cuz we went through it with them last time. And, you know, we knew we knew him and felt pretty comfortable just starting to wire. So that deal came together really quickly.
I did it on mother's day from an iPad and that's how things go these days. Sometimes we have to move. We move really quick. The moment it feels like we have a bit of momentum, you know, we, we have in the past gone several months without buying anything. And then the first three acquisitions we ever did were all within a few months of each other.
Now we just did cold DM a couple months ago, so that one's feeling comfortable. We sort of are, have that thing on an incremental path now. And now we're doing this one, but we have another one under L LOI already. And another target that we're trying to decide if it's, if it's a good fit for us. So it's possible that we could do three in the next seven days or two additional one in the next seven days. So the point of support guy, I, you know, I don't want to get too much into the product, but it's one of these tools where you can upload.
a your documentation and it will index it in a particular way that allows us to use chat G P T or the G P T API. To enable you to ask questions and have the answers be informed by the documents that you upload. So there's a a ton of these tools and the reason we like this one is there's plenty of horizontal ones. Like we're investigating some of those, like chat with your PDF or chat PDF or.
There's, there's a ton of these tools where you can upload a PDF and then ask it questions about it. And it's a really, really awesome use case. The problem I see with it now that I I think that is preventing us from moving forward with one of those options. And one of those is on the table for us is that the pricing sucks. Everybody's charging like it's either free or five bucks a month. It's just peanuts.
And, and this was cool because it's verticalized. So this is a particular use case, right? If. These, those same guys that were charging five bucks a month, went to a law firm and said, we will help you improve your discovery by 20%, right? When they go into discovery and they get dumped a bunch of documents.
And they have to sift through it and kind of sort things out and figure out what's going on in those documents. If they could just ask intelligent questions about it. I bet that would be pretty interesting for them. They're obviously not gonna pay $5 a month for that. Right. So anytime that you can go vertical, especially with this stuff where there's a ton of competition coming down the line, I think that that's where we're focused It's not so much on on finding some horizontal technology.
But some vertical use case. And even support guide to be Frank is, is a pretty horizontal use case. There's there's a ton of things you can do with chat. at, right? like this, this real estate in the lower right hand corner. That is what this product That's where this product can go. Is really precious real estate. And I don't think a tool like this is quite ready to replace an intercom type tool.
Intercom obviously will. I would, I would expect that they would add something like this, but for plenty of other projects where perhaps there's a few documents on an API and you don't have anybody staffed it's quite nice actually to, instead of just kind of searching and hunting through.
hundreds of articles to just be able to ask the question and get an immediate answer. Without having to search through a bunch of documentation. So, you know, I've been in machine learning for a really long time and mostly doing computer vision stuff. That's, that's kind of my background.
Back in the day. Security and surveillance was like one of the primary use cases of, of computer vision when like the last, and this was, I don't know, 20 15, 20 16. The last time AI was really hot. And. You know, there was a lot of hype around computer vision and all these computer vision startups kind of came and and a lot of them went several, you know, fairly sizable businesses were built in the computer vision space, doing all kinds of cool stuff.
But a lot of 'em went to zero. I I think this time though, it's it's different for me. I think anybody that's tried these tools or done anything with chat G p t has realized just how valuable it is and just how immediately valuable it is right There's there's not a lot of. There's not a lot of guessing about, oh, well, you know, this is really cool tech, like, you know, blockchain stuff like, oh, in theory, it's so cool. So cool. But like, what do I, what do I use this for? I think there's 1,000,001 use cases for this stuff immediately.
So this will be our entry point into, I think, a flurry of activity for us in this area. I still, in terms of our thesis think there's kind of two paths and this thing falls into. Two categories for us. The first is this idea of being roughly vertically integrated. I don't know if I like that word very much, but the idea that we, as an acquirer of small SaaS companies, we buy a bunch of tools to run these things, right. We buy Intercom, we have a help desk. We have a knowledge base.
We buy notion, slack, all these other tools we're using Stripe, right? We have these like little feedback things. We have bug boards and, you know, places where people can upload their thoughts about the product net, promoter scores, banners, announcements, all that stuff. There's kind of a fun concept. That's, that's sort of starting to congeal where if we continue down this path, it makes sense.
If we continue down this path of buying more companies, it makes sense for us to own some of these key tools that are either a really expensive. Or b just. Necessary across the entire portfolio. It starts to make sense for us to acquire something and sort of eat our own dog food and be be a customer of the tool.
And then also sell that tool as a SaaS. And I think that that's a really interesting. Concept It might help us sort of frame instead of being just purely opportunistic with our acquisitions, it can help us frame a little bit better. What we're trying to buy and why, why something might be a yes.
And why something might be a no just on a product level for being a good fit for XO. So this sort of fits into the vertical integration thesis. Every single one of our tools has one of these support chats. And so we're gonna play around and experiment and start installing this and seeing what people say, what people like this one was a really cheap one with no revenue.
Or a minimal revenue. So this is, this is a, a relatively cheap experiment for us to just get our, get our feet wet in this space and figure out if there's something interesting here. I think there's a lot of engineering work that we could do. And so it's gonna be, it's gonna be about us, listening to customers and figuring out how to verticalize this thing even more, but already today.
It's pretty cool. You can embed it as a chat widget or share it as a webpage, so it can just be kind of a standalone webpage. So. You could think that you could see this replacing docs.whatever.com. you just open the doc site and it's this chat widget. And you just ask it the thing that you want to know.
And it just gives you the answer, right. Instead of just having to navigate through all the various different ways that, that people like to do documentation sites. So the second, the second category of why this makes sense for XO is I've already talked about this a little bit, just betting on, on AI.
And you know, AI's a, a really broad term. I think for us, that means in this case, or in these cases, large language model enabled tools which support guy is. So I don't know that we'd go out and buy a computer vision application at the moment, but something that we can build an AI first experience around using, using a large language model. Like.
Like open ai. That's where we are going to make a couple bets. I don't know that all of them are gonna work out. In fact, I expect most will not. But we're sort of, you know, as we think about what a thesis is for X Excel and, And what really is, is the best targets for us. We're still absolutely looking at.
You know, something like Construction software for some, some particular industry, right? We're still looking at those types of tools. But we are also making these kind of bets. That could be a lot bigger. And and grow a lot faster. But the outcome is less likely to be. Non-zero. I said that backwards, these small bets like support guy, if they go to zero, that is the expectation for eight out of 10 of these is that they will actually go to zero. They won't get product market fit.
Or somebody just comes out and says, they've raised a whole bunch of money to go after this exact use case, but there is a vacuum right now for, and and a lot of appetite for trying these new tools. And so we're gonna try and sneak in here and see if we can grab some, some market share because it is a, it is a land grab right now, out there.
So these AI first experiences like this, I think are going to be, I think you'll see a lot more of these types of, of products and the nuance will start to be where an Intercom, for example, not to, not to pick on them, but. Intercom will bolt this technology on to their existing product suite. It will deliver some value, but there's gonna be a whole category of tools. And I don't know exactly what they are yet, where re-imagining the entire product as an AI. First experience will be what eventually beats out the incumbents or, or, you know,
More often what the incumbents will, will acquire. Right. But people will get used to these AI first experiences. And I think once, once they do, and this has certainly been true for me. Once I have that experience of getting an answer immediately, the right answer, without waiting, without talking to anybody or just some, some actual AI first experience that works. It's really hard to go back.
And so that's kind of the bet that we're making here. In terms of the asset transfer. I created another little video, just kind of walking through sort of in real time, how we did the asset transfer. You can watch that separately or listen to that separately. Nothing fancy there. We just throw some stuff passwords usernames into a Google sheet as we're, you know, kind of chatting with.
The, the seller on Twitter. And that's really, that's really it. And then we just kind of make sure. That we have everything that we need before we send the final payment. And then we send the final payment again. This took this took a couple hours on Sunday on mother's day. And, and this was a really quick one, but at least I documented that process. So I'll share that video too.
And I think that that's it. We're really excited to have support guy in the portfolio and stay tuned for acquisitions.
Number is nine and 10, which I think are gonna be happening really, really soon.
