Sachin: what AI changes—and what it does not

Currently a lot of brands, they are using a set of predefined rules to communicate with the consumers. And here the AI comes into play. AI will keep on becoming smarter and smarter. The primary reason for that is the fuel to AI is data. AI can access that data, analyze that data and then create real-time conversations. AI can access that data, data, and data. So the thing is we see brands, businesses using AI-powered predictive analytics to optimize and hyper-personalize email subject lines, timing of sending or using machine learnings to personalize even the content. So if I may start with you, Sanhu, how do you see AI in the space of email marketing and what trends are emerging or what practices are you implementing? Sure.

Email marketing, the most important metric is the delivery rate. How many people get these emails in their inbox? Because if we talk about the market share, Gmail has the largest market share and Google keeps on making it more and more difficult for spammers to lend the email in the inbox, right? So that is the first metric. Second metric is open rate. How many people who received an email, open the email? And I believe AI can help in both these metrics. Because rather than just spraying it and then pray that people don't mark you as spam, we can try to predict that if I'll send this email to this particular guy, the chances of marking this email is way higher than someone else.

Sometimes the timing on which this email is sent, because if you have your notification on, you are in a meeting, suddenly a mail pops up, it's a spam or it's a promotional mail, there are very high chances that you will get irritated and mark it as a spam. Rather than if you are sitting, waiting, the airport waiting lounge, you're waiting for your flight and suddenly a mail comes, you have more time in hand, you will browse it, right? So timing is a very important factor. The second thing is the behavior, the consumer on your previous emails. You can't, while you can create a set of rules using the data, but you can't just bend the rules every time some behavior changes, right?

While AI can do that, because it will adapt, it will learn and change the rules, change the delivery of an email basis, the consumer behavior. Second thing is the open rate. While in previous times we used to have three, four options and basis, many platform they used to provide this ability that you give me four lines. I will try and do an A-B testing and whichever subject line is working better, I will try to post emails to that. Again, that's not a very effective way. If I can change my subject line, while there was a name inside, I can address you directly, I can do it at scale, personalization and all. But what if I can change the entire subject line at scale for each individual, right?

It will again improve my chances of getting my mail open and getting higher response rate on my emails. The third thing obviously would be the creative. Now there are AMP email providers in the market where you can have dynamic emails, you can have dynamic sections in the email. Conversational. You can have conversational emails now, so you can run a chatbot within an email body. You can trigger a chatbot, do the entire conversation in the email and then complete the journey then and there. We are using it. There are a few use cases. At Paisa Bazaar, we run a pre-approved program. Where rather than you applying for it, if you need a loan, you come to us, I'll tell you that you have four pre-approved offers from four of the banks.

And the debt of getting money in your account is drastically reduced. So if in a normal application, you go, you give me all your documents, I'll submit it to the bank. They'll work on the application, they'll give you whether it's approved or not. If it's approved, then there would be the KYC formality, then the money will be dispersed in your account. In a pre-approved process, this entire process is reduced to just KYC verification, either through Aadhaar OTP or something else and the money is in your account. For this process, we don't want people to come onto our site, complete the journey, etc.

So we are trying to do this entire journey with an email, trying to tell the customer that these are the offers, let him or her choose the offer, complete the KYC with an email and just disburse in the email. That's very interesting. So this is how we are using. So how do you see AI leveraging or creating more personalized and engaging conversational marketing experiences? And what trend do you see or you're implementing? So before I start on that, I'll just give an example. I saw a tweet a few days back. Someone traveled to an airline and he tweeted that thanks for flying me to Delhi and sending my luggage to Bangalore at the same time. He was being sarcastic, but there was a reply from the brand. Thank you very much for your kind words.

So basically, it's funny, right? Currently, a lot of brands, they're using a set of predefined rules to communicate with the consumers. While the communication may vary, this is the sentiment, this is the mood of the consumer, this is a number of factors. Right? It's difficult for a human being or a set of rules to analyze multiple consumers at the same time at scale individually. And AI can do that. Because every customer, they have different attributes. They give different signals. And we are in an age, like Grant was mentioning, that this is the dumbest AI that will ever be. And the primary reason for that is the fuel to AI is data. And we are generating a lot of data. Everyone is wearing a smartwatch now.

We are generating data as we go, every second. So, AI can access that data, analyze that data, and then create real-time conversations which are very much in context rather than using a predefined set of rules. And it's true for all industries. So, AI can access that data, analyze that data, and then create real-time conversations which are very important. So, AI can access that data, and then create real-time conversations which are very important.