Capital Efficient Strategy for Generative AI Startups: RavenPack CEO Armando Gonzales (Part 3)

Sramana Mitra: Let’s go to your story. How did you build the company?

Armando Gonzales: In the early days, we were mostly doing research and development and trying to figure out how some of our ideas could apply to different problems.

Sramana Mitra: When you got this client who wanted to use your trading technology, did you start zeroing in on that particular use case as the company positioning?

Armando Gonzales: Exactly. It was the example that got traction. It was a problem that we thought we could solve and we were able to prove that there was feasible technology for that and that we had a good approach to dealing with the problem. To quantify that, we were pitching that we could systematically read every article that’s being published about stocks.

Back then, about 20 years ago, that was tens and hundreds of thousands of articles. Many companies back then were saying that they can read the entire web. We were saying that that’s insane. You’re not going to be able to do that.

You need to narrow down the sources that matter. How do you determine the sources that matter? You go to prime sources like Dow Jones or the Wall Street Journal. Grab that content and try to generate signals off of premium content as opposed to mining the web.

Sramana Mitra: The sources that you’ve named are proprietary sources. How does your algorithm get access to that?

Armando Gonzales: We approached the publishers and partnered with them. We license the content from them.

Sramana Mitra: In real time?

Armando Gonzales: Yes. As part of the partnership, we are able to access everything they’re publishing. We focus on the premium ones.

Sramana Mitra: Does that mean that the subscriptions that you’re selling to your customers, a percentage of that is going to the publishers?

Armando Gonzales: Yes.

Sramana Mitra: What were your customers willing to pay for something like this? What did you learn in your validation work?

Armando Gonzales: We were trying to understand what the right price was. We would come up with a few options that are reasonable based on research. One of the theses there was what percentage is the customer willing to pay from the value that they’ll generate. We’re helping them generate signals so they can make a return. Is that a percentage of what they’re able to make?

None of our customers are going to disclose how much they make so you have to use a proxy. They should be able to make at least 10x the money that they’ve made us. Around that estimate, we decided to price it. To make it accessible, we made it based on how fast you wanted content. We turned it into real-time versus daily. Some institutions don’t need real-time content.

We also started partitioning based on how much data they wanted. If you don’t need data for the entire MSCI World Index, that could be a way to partition the product. Our price ranges depending on the different update frequencies and the amount of content that clients wanted.

This segment is part 3 in the series : Capital Efficient Strategy for Generative AI Startups: RavenPack CEO Armando Gonzales
1 2 3 4 5 6