Abhinav Upadhyay on AI and automation and brand trust

Fantastic to have you on the show. Happy 2024. To you as well. But it's an interesting thing, isn't it? So along with choosing your partner, you know, having children, things like choosing college, they're kind of right up there in terms of decisions. But historically, the amount of data actually being used is this big, right? And I think you've touched upon a very, very nerve, you know, nerve cord over here in India. Education is a highly prestigious prized possession in this country. You want to get married, be a graduate. You want to apply to a government job, be a graduate. And this and the dichotomy is that only 27.4 percent of Indians in the right age are actually applying to college.

So the cross enrollment ratio is only 27.4 percent, which means as of now in this country, only 10 percent of people have actually been to college. And the biggest challenge with that is that the large populace, which is now also having the money to send their kids to college, does not have the wherewithal to guide and consult them in making the right choices because they've not been to college. So 70 percent of all audiences that come to us at College Deco is actually coming from tier two, tier three towns and from and from villages. Every 36 state plus sort of union territory, the kids is applying to us. We get about 50 percent of all traffic that applies to colleges in the country. And that's about 11 million people.

In a country like that, to be able to not leverage data despite the talent being available and the technology being available is, I think, Harakiri. So what we're doing right now is actually using data to be able to predict which will be the right college for you based on psychometric tests as well as questionnaires that we roll out to the kids when they reach out to us for counselling. Well, we'll get into that in a second. But you raise a very interesting point, Abhinav, because when you hear data, your mind immediately skips to interpretation of data. But a lot of the issue, of course, that you're solving is families and their aspiring children don't even have any data. True that. It's lack of data that you're solving. OK.

So there's that fundamental thing that you need to give data to people or they'll never make choices. Of course. Talk to us. Go back again to this point about what is it you're doing with data? Because it's if I think about choosing a college, albeit my choice is very binary, it feels very complicated. It's emotional. It's social. It's academic. How on earth can data or the algorithms that you use make at least improve that decision? Let's say may not make it for you. And I think that the key word there is improve and enhance. What we do is augment the decision making process rather than take over the decision making process.

What we've witnessed with the children and you're right, there are cultural context nuances to why somebody or personal nuances to why somebody seeks to go to higher education institution. But there are also family matters. There are sort of multiple financial matters that play a part there as well. So there's rational and then there is the emotional. Human beings at the end of it all. We're not rational people. We're actually slaves of our emotions. Data only contextualizes the emotion and helps the kids make the right choice. Here is how we do it. We understand which cities or villages kids are choosing which course and which stream and looking to go to which part of the country.

So a massive amount of migration happens in India for higher education. Now, in that context, can I help predict that this kid with these kind of marks, with this financial background has a propensity to migrate to a certain kind of city and would be more comfortable being in a certain kind of stream or course in a college of this kind. So we're able to use data analytics and now data science rather to be able to predict that. And once we've done so, we offer the child the choice of being able to look at the options available on the table and thereafter guide the children and the parents as well in making the more informed choices. Just augmentation of decision making with the availability of data. You're kind of you're not a you're not a substitute.

You're a kind of pathfinder supporter brand. Oh, absolutely. If you hit the nail on the head. Actually, we want to be a guide, a friend. So on on the website today, we get about 200 million sessions in a year and out of which about 4.5 million children register for counseling with us. And that's about 50 percent of children who take admission in a year in a college. So pretty much the largest higher education ecosystem in this country today. But what we would truly pride ourselves on being is a trusted friend, a guide to the student who's never going to find the guidance in a tier two, tier three town, not from his family, not from her teachers sometimes and definitely not from career counselors because they don't exist in some of those towns.

So we're able to bridge the gap and the sort of information asymmetry and are able to provide the right guidance to the kids at the right time in a manner and in a language that they understand. So that's also very, very important to us. We don't we don't want to be talking to a child in a language that the kid doesn't understand. But let's talk a bit more about this. So you can see on one level as a brand, you're heavy functional. It's it's a lot of deterministic stuff to help narrow the chances or improve the chances or whatever. But it's very it's an emotional it's a very emotional environment. But I'm going to tell me how do you use data to get a better grip on the emotional drivers of your customers? Very interesting question, actually.

So over the last seven or eight years that the organization has existed, you've you've created enough sort of data points and variables on which that emotional decision making also lies. So we've understood that availability of finance while looking rational is actually an emotional decision. We're able to predict basis monthly household income. What is the kind of streams and basis the grade that the kid has got or the kind of streams that they would be happy with. But we're also using a psychometric analysis to figure out the psychological reasons behind why a certain career or a stream would be the right fit for a child.

And what's also important to understand in making the emotional become more measurable is that we use every possible tool when the child comes on to the page the first time. So right from using a hot jar or clarity, et cetera, et cetera, to figure out what all are the pieces of content that the child is interested in. So what are we serving today? India's got about 54,000 colleges. The United States has got about 6,000. The UK has got about three and a half, 4,000. So in in a multitude of choices, the choice actually becomes tougher. So we offer the information about all 40,000, 50,000 colleges that exist in this country from their exams and grades and et cetera, et cetera. And what would it take to get in there?

But when the child is glancing through that information at the back end, we're also reading what where is the child spending the most time? So I can say that, you know, content is king, you know, but actually data is God. It's helping me understand what content needs to be the king on the page. Well, it's very interesting listening to you, isn't it? Because you're using multiple data methodologies. You're changing and improving them all the time. And no doubt. Absolutely. You use, you know, machine learning and stuff. That's right. To sort of improve. And I'm going to touch upon that in a bit. So it's a kind of it's it's it's a journey. We're getting into the final straits of our mini journey here.

But I want to talk a bit about responsibility, right, because in the same way that this is a massive life choice, your brand responsibility. I mean, you're not in the business of getting it right. But your data drives big life decisions. How do you as a brand and a business think about that awesome responsibility and how do you make good on on being good to it? Yeah, I think in every brand and every company needs to come up with its own modeling around it. But we believe that by staying neutral to the student in terms of the choice that the student makes, she's going to make a choice eventually based on what the narrowed down choices are and not 54,000 colleges.

So can we offer her the right choice for one, the financial background to the stream that she's chosen earlier interest areas or the right careers based on a psychometric test, but largely also what the family would be OK with. So we offer the choices. We stay neutral on the platform. We don't say that you necessarily need to take an admission here because that's where you'll get a seat. We help her apply to multitude of colleges via one single application, common application form, and she eventually makes a choice. That's one aspect of respecting her choice and helping her make the right choice rather. Second is, you know, when we're helping a person make a choice, I am also responsible to protect the information that the person has shared with me.

So do I need to comply to the GDPR norms? So data privacy is very, very, very heavily guarded in this company. We simply don't have enough time, but I just want to end here, right, because you've worked across multiple businesses, a lot of them tech businesses. Why is your current shop any more responsible than any other tech brand? In a word. I think compassion, you know, nation building. And I've worked in companies, not just tech, but also legacy brands. So I started my journey with the Tata Group. I passed it over to Micah and I reached the Tata Group building their telecom business in India. I'm sure it was tech, but it was still a legacy sort of a company. The practice is very legacy despite the tech being modern.

And the Western Union, which is also a legacy brand, 150, 160-year-old company, right? I'm now seeing a shift from being just from storytelling being just an art to storytelling actually being led by database decision making. So I would say compassion towards the people that we work for while being fully accountable to where we are today and what we're looking to do for them. Well said, Abhinav. I love that you can start a conversation with data and end with compassion. You've been a fantastic guest. Thank you so much. It's been fantastic chatting up with you as well, Jasper. Lovely talking to you. Thank you, sir.