1Mby1M Virtual Accelerator AI Investor Forum: David Evans, Sentiero Ventures (Part 3)

Sramana Mitra: Tell me about your AI investment thesis.

David Evans: I’m going to build on something you said in terms of how do you get sticky, how do you know what to invest in? Generally speaking, we’re looking for companies that meet two criteria. One is unilateral where AI is core to the value proposition. If you just have a button that says, “Click here to do something AI,” we’re really not interested. AI features are table stakes.

Here’s how I explain this. Uber is a mobile company. If you take away the mobile app, Uber no longer exists. Amazon is an e-commerce company with a mobile app. If you take away the mobile app, Amazon still continues to do business. They’re not a mobile-first company. To extend the analogy to us as investors, we’re looking for companies where AI is as essential to that company the same way mobile is to Uber. If you take AI out of the equation, there is no company.

A great example of that is one of our companies – Scalestack. They have no user interface. It’s a pure agentic solution that works in the background with the existing tools and products that the company already uses. Nobody logs in to Scalestack to do anything. It’s pure AI, pure agentic backend.

The second thing we look for is one of two things: either you have a data moat or you have an algorithmic moat. That’s how you stay sticky and how you have long-term longevity. You don’t avoid the copycat entirely, but you mitigate it. If you have some dataset that doesn’t exist anywhere, or if you have some level of algorithmic structure or workflow that doesn’t exist or is incredibly difficult to replicate — that’s what we’re looking for.

So, those are really the two things: are you AI-first and AI-native, and do you have a data or algorithmic moat over your competition?

Sramana Mitra: So, let’s double click on some of the investments you’ve made that fit these specs, so to speak.

David Evans: Yes, absolutely. One example of a data moat is a company by the name of SingleKey, which is from our first fund.

SingleKey works with independent landlords to do background screening, rent collection, and offer a unique insurance product. If a tenant breaks the lease or needs to be evicted, they insure against the losses. Because of the process they facilitate, they know what the characteristics of a tenant were when they applied to rent, when they paid, if they needed to be evicted, or if they broke their lease. That dataset doesn’t exist anywhere else for independent landlords. That’s what you can call a data moat.

They’ve been doing this for nearly five years. They’ve got five years’ worth of data that even someone new to the market wouldn’t have. You don’t have that longitudinal data or the varying economic backdrops. That’s a company with a data moat.

When we talk about an algorithmic moat, we have a company called Geminus. Geminus incorporates both the laws of physics as well as the data in training machine learning models. They’ve got a proprietary approach to utilize those two together to require less data and less compute time to achieve not only equivalent but also better results than with just the data model.

So, those are two examples of companies in the portfolio that fit each one of those buckets I described.

This segment is part 3 in the series : 1Mby1M Virtual Accelerator AI Investor Forum: David Evans, Sentiero Ventures
1 2 3 4 5 6