Building a New Age AI Services Company: Cognida CEO Feroze Mohammed (Part 4)

Sramana Mitra: Feroze, there are a few things that come to mind as I’m listening to you. First and foremost, you talked about delivery constraints and supply constraints in terms of the expertise that give you this moat. You haven’t talked about domain knowledge.

A company that I’ve been following for a number of years that recently announced a very big kind of roll-up is Machinify. They had started to do AI services, the same model of IP-led AI services that you’re talking about. They built up a very interesting company. But very early on, one of their top domains was in the health insurance business. That is going to be one of the big differentiators for companies that get built in this moat as you accumulate domain knowledge and domain-specific use case knowledge. That is going to be a very big differentiator.

Feroze Mohammed: Absolutely. A lot of these AI solutions need deep domain expertise to solve for, right? I couldn’t agree more.

Sramana Mitra: To some extent, if I were doing strategy for your company—I did many years of strategy consulting before I went back to being an entrepreneur—I would pick a domain and go deep into a domain.

Feroze Mohammed: Absolutely. I think our initial journey was trying to figure out where we are getting the solution-market fit.

Sramana Mitra: Yes.

Feroze Mohammed: That’s the reason why I mentioned these three industries. I think we do a lot of interesting work in the manufacturing space. We do a lot of work in the healthcare insurance space, particularly in the subrogation business, and in the asset management space in the financial sense.

Sramana Mitra: But these are disparate.

Feroze Mohammed: These are disparate at some point.

Sramana Mitra: They are disparate domains and disparate use cases.

Feroze Mohammed: Eventually, I think we will declare our religion – pick one of them and go much deeper into them. But at this stage, as we are trying to scale, one of the other aspects that we also are hungry about is doing a lot of use cases in this space.

To my knowledge, again, not being boastful, for a company of our size, I think we do some of the most complex use cases in the business. That gives us a lot of understanding. A lot of the capabilities that we’ve built are portable across industries. That’s the reason why we focus on getting as many use cases as possible.

But I do agree with you. We’re kind of narrowing our focus. Eventually, I think we will end up declaring our religion, pick one or two, not just industries, but within the industry, that specific intersection of a process and a sub-process in an industry. I think that’s where we would want to own that particular sub-process in a given industry.

Sramana Mitra: Now, switching gears a little bit, let’s talk about the commercial aspect of these kinds of companies. You very astutely articulated the dynamics of these $10 million contracts that Infosys and Wipro are doing versus these $200K contracts that require very specific kinds of expertise. And you’re willing to do those contracts to learn those use cases and build on them. How do you price? What is the pricing model of your company?

Feroze Mohammed: Yeah, so basically there are three elements to our pricing. One element is initially when we build the solution. So, there’s an initial pricing for crafting or tailoring the solution for the business.

Then we have a runtime cost of an MRR for our IP. The IP is in one of three forms: either it’s a model that is exported out and runs on whatever platform or infrastructure they have; or our entire software runs as a container in the given hyperscaler’s environment; or we run and manage it, including the compute and the storage. So there’s an MRR model for that.

The third component is, apart from the IP, a lot of these solutions continue to require white glove services because data drift happens. The data pipelines need to be managed and maintained. Getting the data ready for AI and continuously keeping it relevant for AI requires a lot of activity. So that again is priced in an MRR model.

So we’ve got the initial consulting piece, depending on the scope of the work. Then we have an MRR for the license. Then we have an MRR for the white glove services.

These two elements – the MRR for the license and the MRR for the white glove services – are non-linear. The pricing for the initial part of the solution is sometimes outcome-based. It’s a fixed price model that we do.

This is where we have landed after a lot of experimentation over a period of time. This seems like the most practical model – easier for clients to understand because there is a defined scope. They know what they’re paying for. They know that they don’t have to break the bank to solve the problem.

At this point in time, this is where we see more of these kinds of constructs, but I’m assuming we will learn as we go along.

This segment is part 4 in the series : Building a New Age AI Services Company: Cognida CEO Feroze Mohammed
1 2 3 4 5 6 7