Building a New Age AI Services Company: Cognida CEO Feroze Mohammed (Part 5)

Sramana Mitra: Right, you’ll learn as you go along. But there will be these three buckets, I think, even when you are a $100M or a $1B company. This is going to be the structure of your company. So the question then is, what is the distribution?

If you look at the revenue, what percentage of it is in the IP bucket? What percentage of it is in the white glove services bucket? What percentage of the revenue is in the setup fee or the consulting setup fee?

Feroze Mohammed: Today, it is about 75%, 15%, and 10%.

Sramana Mitra: Is 75% IP and 15% white glove services?

Feroze Mohammed: 75% is the initial setup fee. About 15% is the MRR for the IP. 10% percent is the white glove services. Again, when I say white glove services, these are not traditional linear services. These are services that are multiplexed across all the many installations that we’ve got. So, it’s 75%, 15%, and 10% at any point of time for a given client. But what you would also need to understand is that this 15% and 10% get compounded over a period of time.

Sramana Mitra: Of course.

Feroze Mohammed: If you were to look at the revenue spread five years down the line, we would probably have reached a revenue spread where these two lines will probably be more than 50% of our revenue. So, per deal, it is structured like 75, 15, and 10. But as we deploy more and more solutions, I think these two components would eventually reach up to more than 50% of the overall revenue.

Sramana Mitra: OK, so ultimately you’re trying to get to more of a SaaS kind of repeatability and predictability in your revenues.

Feroze Mohammed: There is a non-linearity and predictability in the revenue that we would seek. I would say this is different from SaaS in the sense that a lot of this continues to evolve, right? In SaaS, it’s a one-time point solution that you sell. Whereas here, we build a model, then we build an adjacent model, then another model, and so on.

Sramana Mitra: That’s true about SaaS as well. You sell different modules into the same customer base. I don’t think that’s necessarily different. What is different is the white glove services portion is changing as new things happen, new use cases come about, and so forth.

So, I guess my next question on that white glove services piece is, what is the scope for automation and margin extraction in that white glove services portion?

Feroze Mohammed: Significantly high. The entire ML ops, if you were to broadly call it, it’s AI ops or ML ops as the white glove services. Number one, it’s highly multiplexed. You would need the fractional input. Let’s say every time a data drift happens, you would need data science expertise to retune or retrain the model or tune the parameters.

One, it is multiplexed across different installations. Second, more and more automation is happening in that space. The entire ML ops part and the AI ops part of it – right from monitoring of it, figuring out when the change needs to happen, and the actual implementation or changes to that – is also getting more and more automated. So, we sort of eat our own dog food as well. We use a lot of AI in our own product development, of course, for building the solutions and also for running and managing this. Even the running and maintaining part of it will be using more and more AI.

Sramana Mitra: The setup part is probably the lowest margin part of your business currently. Is there scope for a lot of automation and margin extraction on that part of the revenue stream?

Feroze Mohammed: I need to nuance it out a little bit. It’s not a typical setup. It’s a consulting-led model where we actually work with the businesses, understand which particular use case would work for them and ideate. So it’s a consulting-led model and is a high margin model in itself.

Of course, in the initial part of our journey, when we were growing our business, we were hungrier and were pricing it aggressively. But at this point of time, as we have got more and more repeatability and as we are able to show our library of offerings, that particular segment of business itself is a high margin business.

There’s a degree of repeatability there as well. For example, we have built our own launch pad – a service where in about two days, we conduct workshops and actually give two use cases. We sit with the team, ideate, demonstrate what use cases are possible, work with the business, understand, and leave two working use cases with them. That’s been a very popular model. There’s a degree of repeatability that’s coming in the consulting element.

Sramana Mitra: As you develop the use cases, there’s going to be a lot of repeatability.

Feroze Mohammed: That’s right.

Sramana Mitra: Which is the case for doing it within very similar verticals.

Feroze Mohammed: Absolutely. I think coming back to the point that you mentioned, over a period of time, that gives more and more differentiation.

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