The Exchanges

Q Uh, is, is that all the signals that you are looking at to make a yes decision?

A Um, so I think that's a number of it. I mean, the other thing we think a lot about is go to market. You know, can this team execute a go to market that in our view is very relevant for this industry or the right way to scale? Now, it is very hard to know in the early days what works. A lot of it is experimentation. But when we think about sort of the value we can bring, and having seen this pattern across, you know, now hundreds of companies, we tend to spend a lot of time with them on figuring out, all right, listen, You, the founding team, can clearly sell. You know, you have the domain expertise, you have the passion for the space, you have the founder title, um, you know the product better than anyone, but obviously you won't scale beyond a certain point. So how do we take sort of the playbook that is in your head and build a team that can go execute it? So we think a lot about that and the, the potential to go get to that next stage. Um, now then, you know, I think there's general turn things. We, we, in most cases, we like to have a thesis on the market. There are exceptions to that, but we need to have, we, we would like to build a view on where we think the market goes. Um, and then where this team has an advantage that is hard to replicate.

AI assessment note: “the other thing we think a lot about is go to market”

Q And, uh, what's the, the lowest ARR, or, like, one is lowest and the other is average ARR that you usually come in?

A So we have done companies at incubation in founders we know. It is rare. I'm trying to think of an example where we've done it in people we don't know. I think it's unlikely. Um, so beyond that, I would say the average, gosh, the average is a tough number, but I think we've done everything from companies that are at a few 100,000 to a few million, uh, to start off with. Again, I, I don't want to say, like, you have to hit a million ARR. The number also matters on, depends on the team, you know, how strong our thesis is in that market, how competitive it is in the market. Like, I'm just taking a simple example, right? Like, if you are building a new sales automation tool using AI, there's a lot of companies doing sales automation. So there, it's unlikely if you're at a hundred K, we would come in, or let's say a hundred K is too less, but let's say a hundred K, because There's probably a hundred companies doing that, right? So we need to show something, we need to see something that you're building that approaches the market in a different way or solves it for a different buyer or persona before we would invest. So there the bar is a bit higher in terms of traction. But if you're building, you know, there's a, there's a company called, um, Tote, which I mentioned to you before, are called building something for convenience retail. That is a market that every deal is like a seven…

AI assessment note: “we've done everything from companies that are at a few 100,000 to a few million”

Q And do you see in the current AI wave, product market should also changing dynamically for companies?

A Yeah, I mean, it is changing in the sense that the pace at which you can get a buyer and potentially even churn is much higher, right? Because the AI products, one of the benefits of that is it is easier to build, but also I think it is easier to put it in and replace it, right? In some cases, it's harder, like, especially in, like, certain industries or, like, certain products, like, with compliance and all. Maybe that's not a good example, but, but You know, 1015 years ago would be much harder to, like, get a product in, but then ripping it out would be much harder, right? So now, especially if the customers have done a good job, you know, managing their data, um, you know, building models on it, and so forth, and depending on how the, the rights there are assigned, you can get people in and out very, very quickly. So that increases risk a lot. So yes, you can get product market pay very quickly, But the next product comes along, and they can deploy quickly as well, which is why when we talked about earlier, I think companies really need to think about how are you building that full stack between the model layer and the UI. Like, you need to own that entire thing, because if you own the entire thing, at least in some verticals, It is a lot of work to rip you out. And second, if you build that well, as the models get better, your product just gets better. So somebody comes in …

AI assessment note: “it is changing in the sense that the pace at which you can get a buyer”

Q Are you also thinking of, uh, let's say more hardware develop, getting developed, so AI for, uh, AutoCAD kind of things, is that also?

A Yeah, I mean, I think, listen, any software that follows Roughly a set of rules and is algorithmic, I think AI is a great use case for that, right? So if you're doing design or CAD where, you know, there's a set of best principles, I think AI does that well. Anywhere that requires a lot of human judgment or, you know, design thinking or creativity, I think AI will get there reasonably well, or maybe the, it can do a good job if you have no creativity or skills. But then that's where it becomes much harder, right? So if you are in an assembly line, I'm just using that as a very general example, and you are, it's, it's a very repetitive task. That's where the AI can get good at because it's repetitive. You can train it tens of millions of times very quickly.

AI assessment note: “if you're doing design or CAD where, you know, there's a set of best principles”

Q So how many of these companies you would, you would do?

A Yeah, so I think in, I mean, a significant chunk of our portfolio, I don't know the exact percentage again, is, uh, founders of Indian knowledge, right? Many of them were born and raised in India. We have some that were born and raised here. Um, and, um, I think in most cases, the, at least one founder is here. Almost everyone has teams in India. Um, so, um, We are totally fine if most of the team is there. I think in general, you have to be in the market you're selling in. So it is rare, I think, that a team can be fully based, let's say in India, and then sell into the US. You can maybe for some kind of products, but in general, if you, as you get to like mid-size and enterprise, like they want to meet you, you need to hire people here, and so forth. So I think in most cases, There is some presence here. I think that's not just true for India. We invest globally. Even other countries, typically a founder moves here.

AI assessment note: “a significant chunk of our portfolio, I don't know the exact percentage again”

Q Understood. And let's say, uh, out of the companies that you, uh, invest in, like how many, uh, cold reach outs would you, would you?

A So cold reach outs is tough. Uh, honestly, just because we get so many, um, and I know it, it's probably not fair, but the reality is, Gosh, I don't know. I maybe get 10 to 15 cold emails a day. And this doesn't even include LinkedIn. So I'm, if I haven't responded, you know, I'm sorry. It's just that I just, unfortunately, sometimes just don't have time to read everything. Um, and then, but this, this could be helpful advice. So a lot of cold reach out emails tend to be very generic. Hey, I'm building something for this. Can we meet? Right? Those are just hard to follow up on because you only have X number of days and hours in a week, and if I respond to all of them, I won't have time for anything. I think crafting a very good cold email which says, you know, I am so-and-so. This is my background. This is the problem that I'm solving, and I think this needs to be solved for this reason. Like, this, this is a massive urgent pain, and then this is our growth rate. We have customers. We have grown quickly. Um, I think that is much easier to process and decide whether to follow up on or not. Um, I'm still probably not gonna respond to every cold email, but at least I, at that point, can make a quick decision and, you know, at least respond with, sorry, this is not a fit or not. So maybe that's, that's advice is, you know, we've definitely done deals in the past where You know, in …

AI assessment note: “we've done cold deals, but like, it's rarer”