Nimish Agarwal: what AI changes—and what it does not

Health insurance is a complicated category. It's not a lifestyle category. Lead time for a consumer to get into the journey and actually make a purchase is anywhere between two months to two and a half. There are very large call centers we don't have. We've got a small call center that we've created a chatbot. Of course, chatbots have been around for a while and multimodal communication. But companies are finding innovative ways to use them in conversational marketing. H&M's kick chatbot or we've got even Tuck and HDFC, the KEA and the EVA. I like EVA more than KEA. Don't judge me. And yeah, I'm a customer with both. But this thing, so one such example is Sheffora's virtual beauty assistant. Which understands the needs, preferences and response.

What do you think, any examples or any use cases, what do you think could make successful AI powered chatbot which are both that have been launched recently, can be effective? What's your take on the whole chatbot and AI school? Thank you Vatsal. Sitting next to you is like sitting next to a headmaster. My god. Sack. I'll be very measured in what I say. For the group, hi, I'm Nimesh. I lead marketing for Neewa Bhupast while Max Bhupai also happen to lead the digital sales business. On the chatbot stuff, a lot of people, I don't need to preach to the converted. I'm sure I'll pick up a lot of stuff from you. Just a quick example from our category. People don't understand it.

The lead time for a consumer to get into the journey and actually make a purchase is anywhere between two months, two and a half months. That's a very long time, right? And typically, in a Paisa Bazaar, Policy Bazaar, we were very closely with Policy Bazaar. Policy Bazaar has been given to people. What they do is that they fill a lead and then they sit. Because they know in 10 minutes somebody is going to call you. Right? They have a very large call center. We don't have, we've got a small call center. So what we said is that the lot of time, the telecaller has to explain the category to customer. It could be done through a bot. So what we've done and what we're experimenting with is that we've created a chatbot.

Of course, the way the business model works, it's not completely digital. There is an assistance that is required for purchase. But a lot of questions around, I have a group health insurance. Should I take a personal health insurance? How do I compare different health insurance policies? What's the right kind of loan? What is what if I have cancer? And also some very interesting use cases in terms of medical history. It's very personal. Somebody talked about, Sukrit talked about data privacy. We have a very active regulator in this industry. So a lot of conversation with the bot is personalized. I have a certain kind of an ailment. Will I get a, will I get a policy?

Now what has happened is that with that small experiment that we're doing, a lot of efficiency is coming from a time sort of it. Consumers, we also tried selling through the bot by sending a link, etc. The conversion rate was not great, to be very honest. And also the drop offs were very high. And hence what we said is that in the last journey, we ask a question through the bot, do you need an assistance now? And the moment they say yes, now what happens is, we have a finite telecalling team, right? And we are breaking even and we want to scale up our bottom line as well. So we are very cognizant of pushing only those people like somebody, the gentleman from Magic Bricks talked about window shopping.

So we want to keep the window shopping to an intelligent conversation between a machine and a human. But when the customer is really ready, and he's just not trying to understand the category through human interaction, we push him to a human. So that's one thing that we're doing. So