Bootstrapping a New Age AI Services Venture: Anji Maram, CEO of CriticalRiver (Part 6)

Sramana Mitra: A lot of enterprise conversational AI applications that you’re talking about are happening more on the customer end – customer service, sales, or customer relationship management. It’s not much on the finance end. So, you rightly point out that this is a more of an open area where you’re starting to see traction.

What is the business model of how you are going to market with your AI solutions? When you do this conversational AI or automation projects, how are you pricing it?

Anji Maram: It’s an interesting question. We are exploring that now. We are going with some kind of a pricing model that may or may not work. We’re talking to various customers and then figuring out what is the pricing model that works.

Even though we are bringing in a data solutions or AI framework, we’re going to narrow down to an ARR model. At this moment, we are figuring out the pricing. We’re being very flexible with initial customers to help us understand what is the right pricing that we should charge.

Let’s say we already have two customers, and we have a good amount of pipeline that we were able to generate in recent times when we started talking to them. The problem statements we’re able to solve for large organizations are slightly different for mid-level organizations.

It’s all within finance, but their use cases are slightly different, which we are realizing now. Now, how do we train for those scenarios and then make it work for mid-level organizations? How much can a large-level organization pay? How much can a mid-level organization pay?

So, honestly speaking, we haven’t figured out what is that the pricing model that we’ll follow.

We’re going to have an ARR model. To bring data together for an organization, a lot of services companies do data warehouse projects. They charge anywhere from $1.5 million to maybe $4 million, to just bring data together. After that, the customer has to manage the data warehouse and the pipelines. They must continue getting the data, and in addition to that, they need to build reports. They need to build AI on top of that.

We’re giving a pre-packaged data model, which is already 80% ready. Data is different for every customer, but our goal is to have a 80% data model ready for any customer.

Sramana Mitra: You’re tech stack ready with the building blocks of such a model. Then you customize it to the requirements of the customers that you have. Is that right?

Anji Maram: I’m talking about the data model itself, not about tech stack. If you take a financial model, you can pre-build around 70-80%, and there’s an additional 20% professional services work we need to do when it comes to data. Then on the AI, we have a tech stack ready, and we’re training for at least 70-80% of scenarios.

Then, each customer has another 20% scenarios, and each customer has another 20% extra data where we need to build modeling on top of that. Our goal is to take a 75-80% pre-packaged solution that works, and then we go to the customer and say, “Hey, if it takes one and a half year to do this, we are going to enable it for four months. We will build your delta data models and train your delta scenarios, and then make it live for you in four months. We are going to charge much less in year one, and then every year we’re going to charge certain amount for managing this data and the pipelines and then training some of the similar scenarios that you incur year to year.”

That’s the way we are going to price, basically. It will be much lesser than a one-time building cost, whereas we’re going to bank on making it affordable to build the initial data warehouse in the year one. Then after that, we’d charge annual recurring revenue based on whether the customer can treat it as IP-based revenue or support-based revenue.

At this moment, we want to charge a one-time, yearly fee for the IP plus…

Sramana Mitra: Set up?

Anji Maram: Set up basically. That’s the approach we are taking. I think it would work, but we’re nimble enough to adjust our pricing model as we learn more, right?

Sramana Mitra: So, I think the core nugget of what you said is that a large component of your revenue in the future is going to be ARR based because of how you are architecting the company and the business.

Anji Maram: I think I just want to make one distinction at this moment. Our services are still strong, and we are continuously doing services. A year or two years later or after a certain time, we may take this product in a different direction. We don’t know yet, but we will continue doing our application services.

Other than AI, there are still a lot of data services we need to provide- building dashboards, building data warehouses. At the same time, we have an IP that we built, and now we’re trying to figure out how we can price it and take it to market. As we figure out the models, maybe a year later, I will have more clarity in terms of how we take product versus services.

Sramana Mitra: So far, your business is a pure services business. There is no ARR business in that part of the business. The ARR business is starting to kick in gear as you’re introducing AI and you are starting to explore the ARR model in the context of your AI work.

Anji Maram: That’s right. Our prestigious board members and directors like Raju Reddy garu would be helping us to see how we can steer this to the next level. I believe that if we are doing the right things, the next part will be found automatically because we’ve people like Raju Reddy garu to advise us. Our goal is to make it happen, execute it well, and then the rest will automatically follow.

This segment is part 6 in the series : Bootstrapping a New Age AI Services Venture: Anji Maram, CEO of CriticalRiver
1 2 3 4 5 6 7