Building Two Capital-Efficient AI Companies: Arijit Sengupta, Founder and CEO of Aible and BeyondCore (Part 5)

Sramana Mitra: What was your lock-in period? Did you go to Salesforce?

Arijit Sengupta: I had a three-year resting and two-year non-compete clause. I did my two years. Salesforce is a wonderful place. Even if you leave Salesforce, you’re part of the family. I’m an entrepreneur at heart. I wanted to solve this problem and started Aible.

Sramana Mitra: What was the concept of Aible?

Arijit Sengupta: One of the things I was getting into after having had the time to sit back and think was that there was something fundamentally wrong with AI. I couldn’t quite figure out what was wrong. I was going to start a company where I’m going to fundamentally rethink AI.

What we’ve come to realize is, all AI in the world today is trained on the wrong thing and actively harms value. AI today is trained on academic metrics like accuracy, precision, and recall. When I tell you something, how often am I correct?

The problem is, businesses are not trying to optimize that. What businesses are trying to do is optimize profit or minimize cost. If I tell you that the value of a customer is $100 and the cost of pursuing a customer, whether or not they’ll buy, is $1. What advice would you give the entrepreneur? Try hard. Try 98 customers until you get the 99th customer to buy. But that is an incredibly inaccurate strategy.

Imagine you have a 10% conversion rate. What happens in traditional AI is they look at the additional 10% conversion rate and say, “If I tell you this customer will buy, I have nine ways of getting it wrong. If I tell you the customer won’t buy, I have one way of getting it wrong.” The more accurate model is typically the more conservative model.

Aible says, “Rather than optimizing for that, why don’t I ask the customer what’s the benefit and cost of pursuing a deal and not getting it? Let’s ask you your cost-benefit tradeoffs. Let’s ask you how many salespeople you have or how much marketing budget you have. Now, what if I optimize for your business outcomes rather than for academic measures?”

Rather than searching for theoretical stuff, you’re searching for money at a very practical level.

Sramana Mitra: How do you go to market with this value proposition?

Ajirit Sengupta: I’ll tell you that in one second. The AI tells you to pursue more customers. It will get more conservative if the situation were different. So you end up with a lot more money with the same data.

There are two paths to go to market. How do you break through with your message? How do you actually get money from customers? We started this campaign of, “Impact, not accuracy.” I’ve written a book called “AI is a Waste of Money.” We were provoking people.

We did the initial public release where there’s $1,000 per user per year. People tried. We had customers find $3 million of value in two hours. I am not kidding you. That customer is happy to take a video conference. We have videos of Charlie Merrill.

Gartner ended up doing a case study with him where it was Google, Data Robot, HQ.ai, and Aible. These were the four vendors in the case study. This customer is talking about how they found $3 million in two hours. Once you get these early customers who have real stories of how this impacted them, you go to other people with it.

Sramana Mitra: How are you charging these customers?

Arijit Sengupta: $1,000 per user per year. For the second stage, we started with a 30-day free trial. What starts happening is most of these guys have tried creating models before. They know how much time it takes. That’s still the pilot. The bestselling is when you’re taking away from the customer. Then the customer will fight you to the nail to not take it away. 

Sramana Mitra: They’re already hooked.

This segment is part 5 in the series : Building Two Capital-Efficient AI Companies: Arijit Sengupta, Founder and CEO of Aible and BeyondCore
1 2 3 4 5 6 7