The Exchanges

Q And to improve that number, they, you know, focus on tier one colleges only, right?

A Correct. It, it works in, uh, two ways. Uh, a lot of the loans go in tier one colleges where there are almost zero NPAs and then a bunch of loans on tier two, tier three, where they cannot deny because education is also a sensitive, uh, category, politically sensitive category. It's a priority sector lending category. They have to fill up their priority sector, uh, targets also. Uh, there are significant NPAs in those particular segments. For us, the market is in the tier two, tier three segments, which are good enough for us to lend to. Highly underpenetrated right now, and sort of opportunities which are traditional, you know, banks would not be able to cater to, and, you know, we get better over time in underwriting these, uh, institutes and courses.

AI assessment note: “Correct. It, it works in, uh, two ways.”

Q But still the numbers are different, right? Like the, if we just talk about the loans that, uh, the kind of loans are granted for home, for vehicle versus the number of loans that are granted for education. There's huge gap.

A Yeah. So education is categorized as a PSL. Uh, priority sector lending, and, uh, there are targets for that, but overall, I agree with you. Say, for example, uh, housing category or a vehicle loan category, uh, 50%, uh, would be 50% penetrated in terms, as in, 50% of the spends would be financed with loans. Uh, so it's highly commoditized. Now, education is The penetration for education spends would be close to five percent, as in five percent of, uh, loans in overall education as spends, which is close to around a hundred billion dollars. A lot of people are changing that, uh, you know, some of our peers like Anavance and Credila have done some wonderful work in, uh, you know, study abroad segment where it's the same story, but, you know, students going abroad, uh, they've done that and they've proven that particular market too, that They've proven that thesis, which you sort of spoke about.

AI assessment note: “overall, I agree with you. Say, for example, housing category... 50%... education... five percent”

Q How many we would have in tier one? How many tier two?

A Um, tier one would be say close to around 70 or 80 institutes. Um, tier Two would be close to around 302 50 to 300 institutes broadly, and tier three would be, say, around thousand, 1500, somewhere around that, and tier four and five would be enormous sort of institutes. I'm not even going there. So we've had, uh, Experience of working directly on the ground with tier two and tier three institutes, and what we see there is that placements happen. Uh, you know, uh, students are motivated. These are good students. They do something with their lives. It's, it's not that tier two and tier three student pass outs are the ones who, uh, end up being unnamed, unemployable sort of youth. Uh, either they get placed through the Institute and typically we would say, you know, say close to around 50, 60% sort of placement happening, uh, through the Institute and, you know, Anywhere from a four lakh to six lakh sort of salary CTCs, which are good to begin with. And then these students figure out their own careers. Some of them move and get trained into say certification programs or some finishing schools before starting up, uh, you know, small time sort of jobs. Some of them get into sales. But if you look at these students say, you know, one year after having passed out, So, 80, 90% of them, 80 to 90% of them would be in a job and would be doing something or another, would be earning money.…

AI assessment note: “tier one would be say close to around 70 or 80 institutes.”

Q Can you simplify this part? I'm not able to understand like how post LLM world changes things in this sector.

A Oh, it changes like crazy. You know, I, If I were a person starting an NBFC right now, I would start thinking of a NBFC, you know, sort of a ten-person NBFC or a fifty-person NBFC instead of a four-hundred-member team that we have right now. Even that is crazy efficient, as I told you, right, as an, uh, as I sort of mentioned earlier, so we have, what, four-hundred-member team doing one, one and a half lakh sort of loans. But, uh, yeah, as in every facet of, a lot of facets of lending are being, uh, you know, can be very AI first. Uh, you have an opportunity of building those playbooks. So say for example, uh, the core heart of lending, which is around underwriting students, which is credit and risk. Now these models will only become better over time. We've, uh, a weekend project helps us, uh, build, uh, the crystal score of institutes in a much more

AI assessment note: “I would start thinking of a NBFC, you know, sort of a ten-person NBFC”

Q So why, why not banks are doing it?

A I think it's just a question of focus and, uh, you know, we exist as propelled for the last 10 years to solve a hard problem and we are, you know, scratching the surface of it with You know, a post LLM world makes it feel like we are so much empowered now, but the learnings that we have had over the last few years are not transferable. Uh, you know, these are ingrained deep philosophical sort of, uh, learnings and a lot more tactical learnings as well for a bank. It doesn't matter, right? It doesn't the retail book itself for lending. Is fraction of the wholesale book. Within the retail book, Something which has been going on and, you know, secured products are a lot more better. It's been going on. So as an example, in FI-Five, SBI would have given out some 2.25 trillion rupees of loans for home loan. Whereas for education, it would be close to .15 trillion. So 50 next difference. So who in there would want to really focus on this and really crack this and really sort of do this? Uh, that's a question.

AI assessment note: “I think it's just a question of focus”

Q But everything happens in future. So how do you, you know, predict that, uh, this student will be able to pay?

A As, uh, you know, as we have learned traditionally, like the past is a good reflection of the future. It's about doing that and just getting deeper in there. So, um, What we do, you know, both philosophically and tactically in our underwriting and our thought process of risk is get deep into the institutes and really understand what is the quality of service that they provide and how is that going to look like in the future too? Is the student going to get a value out of this? The value need not be just a Just a job. The value could be even something as simple as being able to complete his course and time. Uh, that's a very basic expectation that somebody has once they enroll their kids in a coaching Institute as an example. Uh, so you can imagine that we are building a Sibyl score, uh, or a crystal score for institutes. And for courses. Uh, in fact, I would correct it. It's not for institutes and courses. It's for end use. Uh, and of course institutes and courses are a part of it. It's for a particular end use.

AI assessment note: “What we do... in our underwriting... is get deep into the institutes”