Thought Leaders in Artificial Intelligence: Muddu Sudhakar, CEO of Aisera (Part 2)

Sramana Mitra: Can you go back to the beginning of your journey and tell me about how you got this off the ground? Where did you get the data to train from? How did you get the customers to bet on your system in the very early stages?

Muddu Sudhakar: It was late 2017. The first thing to do is identify what you have to build. Having to stick with it is the key problem for us. You always need to have an initial set of customers. My initial set of customers were companies like McAfee and Zoom.

Sramana Mitra: How did you get those customers?

Muddu Sudhakar: Zoom is always about experience. Eric’s point to me was, “If the users want to ask a question, they don’t need to talk to humans.”

Sramana Mitra: Was he willing to bet on you because you knew him?

Muddu Sudhakar: I never talked to him before this whole process. It all came through his people.

Sramana Mitra: A lot of people are trying to solve these pain points. The question is of betting on you, what did you sell that gave them the confidence to put their data on your system?

Muddu Sudhakar: It’s always the product. We are in high-tech business. The best way to sell a product is for the product to sell itself.

Sramana Mitra: You had the product already before you were talking to them?

Muddu Sudhakar: Yes.

Sramana Mitra: How did you get that data to train that product?

Muddu Sudhakar: Data is very important but you have to come up with the idea first and the algorithm. They don’t have to bet on you. They’ll give you their data. They ran a POC with us. Data was not the biggest problem.

Sramana Mitra: What you’re saying is a different perspective than what most AI entrepreneurs come to us with. They need datasets. Without customer data, they can’t set up the learning system.

Muddu Sudhakar: If someone gives you their dataset, never say no to it. My point is you have enough datasets publicly available. I can crawl Zoom content. It’s publicly available. I don’t need any dataset that Zoom is keeping behind the firewall.

Sramana Mitra: You crawled from the community sites. The public community forums are enough to train your algorithms.

Muddu Sudhakar: To start. One day one, there is enough dataset in public domains for any industry to start. Your results may not be that good, but that’s a good starting point. If you can show results with that, it can only get better as you enrich with more data.

Sramana Mitra: In your case, that is true because there are these community forums that you can crawl. That gives you a starting point.

Muddu Sudhakar: You take McAfee. McAfee documentation is always available on the webpage. We started with pure knowledge. For some customers, we started with only their customer support cases. In some cases, it was sales and marketing.

Sramana Mitra: You don’t only do customer support; you also do sales and marketing.

Muddu Sudhakar: To start with, we were doing customers service, IT helpdesk, lead generation, customer intelligence. We go into other domains like finance and legal. Zoom has all these domains with us.

Sramana Mitra: What did you go to market with? What was the first of those functional areas that you went to market with?

Muddu Sudhakar: In the early days, we went with customer service and helpdesk. For the whole of 2019 and early parts of 2018, we were building the product. We started selling in 2019.

Sramana Mitra: From an ROI point of view, what do you see? We had this conversation earlier about reducing the number of tickets necessary. What kind of metrics are you coming out with by applying AI and your technology to these use cases?

Muddu Sudhakar: You have to have a need for users to improve their churn, sentiment, and other metrics. If that is not their goal, even if there’s ROI, nobody’s going to buy it. A key thing we did for some of the customers was that we improved their metrics by 50 to 70 points. If the experience goes down, nobody will deploy them. That has to be there. Business owners will also want to see ROI. In a sense, is there a zero-sum game here?

For customers in this area, there is a very good ROI. How much is a ticket costing you? That has to go down drastically. Because we are deflecting so many requests, the savings are coming from SaaS licenses and tools. Depending on the size of your company, savings can go as high as $10 million.

Second is, you don’t need to hire for common queries. Some of those can be done by us. It won’t go to a human being. The numbers add up a lot on resources, SaaS licenses, and user productivity.

Aisera also does workflow automation. Let’s say McAfee is not installing on your laptop. Our AI can login to your laptop and solve the problem without a human. I’m using a robotic process to automate fixing the problem.

Sramana Mitra: Really? That is significant.

Muddu Sudhakar: Most people stayed in the chatbot world. We took this to full workflow automation.

This segment is part 2 in the series : Thought Leaders in Artificial Intelligence: Muddu Sudhakar, CEO of Aisera
1 2 3