Podcast #23: How India’s amazing new financial infrastructure will enable better, more inclusive “Bharat UX”Podcast #23: How India’s amazing new financial infrastructure will enable better, more inclusive “Bharat UX”

[podcast]

 

'Banking and financial services 1.0 was you having to go to the bank. Banking 2.0 was essentially the bank came to your phone but you had to go to the banks application. We believe that banking and financial services 3.0 is essentially where the bank is available in every application. If I'm paying off an invoice, I might need credit and therefore loan offers should be readily available at the point of checkout at that time, like it happens in EMI based checkout'

Why do most of the financial products operate on a monthly schedule, when less than 10% of Indians get a monthly salary?

India's household debt to GDP ratio is just 11% (global average ~ 60%). Is it because Indian households do not borrow much?

On today's show, Deepak Shenoy (CEO) and Sahil Kini (Co-founder and CEO, Setu) talk about the problems ailing the Indian financial infrastructure and how APIs are shaping the way digital businesses are built.

Transcript:

(The transcript has been sourced from a premium transcription provider. Yet, it may contain inaccuracies.)

Deepak: Welcome to the Capitalmind podcast once again, ladies and gentlemen, this is Deepak Shenoy from Capitalmind bringing you another fantastic edition. We hope is going to be fantastic because I have with me here a person who is fantastic. Welcome Sahil Kini to the show. He is co-founder and CEO at SETU which provides API for banking infrastructure, connecting financial institutions to all kinds of players who want to use those financial institutions, provide financial services to their apps. So welcome to the show Sahil.

Sahil: Thank you Deepak. Thank you so much for having me. It's such a pleasure.

Deepak: It is indeed. In fact the pleasure’s all mine. And let me go straight to the point, financial infrastructure, API banking, what is it about Indian financial infrastructure that is somehow more attractive or different or in any way superior perhaps to the rest of the world and somehow in my own view, I may be biased because I'm living in India but apparently there's something about India that is much more different and perhaps more attractive. So as a person who's like been deeply involved in this, what really is different about India?

Sahil: That's a great question and a great place to start. I think one of the things to note is that until about 10 years ago, it was very hard to make the claim that financial infrastructure in India is that much ahead or leaps and bounds ahead of what is available in the rest of the world. But basically around 2010 something very interesting started to happen, which is Aadhaar started to take off and Aadhaar was really the foundational layer for identity and identity verification. But it also kind of incubated this kind of thinking of building public digital platforms to basically enable what we refer to as presence-less, paperless, cashless transactions. The first layer of course was presence-less where you verify that you are who you are without necessarily having to physically kind of go to a branch or physically go to a representative.

Sahil: That was what Aadhaar enabled and that's what got extended through electronic KYC, eKYC has gone through its many ups and downs and we believe that it'll make like a full comeback at some point in the future, but that was just the beginning. The idea always was to build a series of layers on top of that digital infrastructure for identity to enable both payment transactions and a free flow of data with the consent of the end user to essentially enable financial products and to enable all kinds of transactions digitally. Obviously financial services was the first focus area because that's the most sophisticated of all the sectors that you talk about when you compare it to say healthcare or your compare it to education, the starting point for financial infrastructure was already further along.

Sahil: And then in 2015-2016 UPI came along and that was really the first major quantum leap after the electronic KYC came along where for the first time you were able to send money from any bank account to any other bank account from a third party application and that's what's really remarkable about it, right? You can dip into your bank account from a phone pay or Google pay and pay to anybody else. That was one main leap that we made which was pure interoperability, like true interoperability between regulated financial institutions, particularly banks in this case. There were several other innovations that people actually haven't even kind of realized the power of which is UPI did something very special, which is it unbundled the addressing layer from the accounts layer.

Sahil: So you can actually have one UPI ID linked to many bank accounts and you can change the bank account, right? At any given point in time the UPI ID has to be linked to one bank account, but you can change the bank account that is underlying your UPI ID. Similarly, you can have multiple UPI IDs linked to the same bank account. This unbundling actually unlocks a bunch of use cases and we can talk about what those use cases are later on, but it's a fairly fundamental shift. The other major shift that it did was it also made authentication schemes uniform across the bank. The UPI pin is a common authentication infrastructure that all the participants in the UPI ecosystem essentially use. It's basically what Visa would have been today if it were to kind of build the infrastructure all over again from the ground up, right?

Sahil: So it's a 21st century infrastructure for payments. And that's been recognized by the financial sector pundits worldwide. And in indeed anybody who's operated in any other economy comes and sees UPI and goes, “Oh my God, this is leaps and bounds ahead,” because it's instantaneous, it is interoperable, it has common authentication infrastructure and it has addressing infrastructure that allows you to do very cool things like generate a unique UPI ID for say every bill, right, which is something that is very hard to do when it comes to bank accounts. That's one major kind of step forward. We've also gone further ahead and said what we did for money with UPI, with payments with UPI, we want to do for data as well. And this is where the NBFC account aggregator framework, which the RBI actually notified in September of 2016 if I'm not wrong, and that's been gaining momentum.

Sahil: It hasn't had its UPI moment yet, but already we have a few regulated entities that have received an in principle approval to become the consent manager. And what the account aggregator is, is essentially for financial data, what UPI was for payments. So it's a protocol for consented data flows from any financial information provider to a financial information user. And what that means is listeners of your podcast definitely would have heard of Plaid which was recently acquired by Visa for $5.6 billion. And what was special about Plaid, right? Plaid essentially for all intents and purposes was really like a Yodlee or a Perfios in some sense in the sense that it built connections to all of these banks and if any application wanted your banking transaction data, you would login to that bank account through that third party application through applied SDK and then give permission for this information to be fetched from that account.

Sahil: What the account aggregator framework is Plaid as a protocol, right? It's a protocol that standardizes the data fetch and data transfer and the encryption end to end that needs to happen for financial information to flow from the institutions to an application that wants to use that information.

Deepak: So you avoid the whole screen scraping and reading of the HTML versus using an API, a proper API to do this.

Sahil: And a much more secure way of doing so because ideally you shouldn't have to at least … so Plaid was one step above screen scraping because it did not use bank credentials. You are to use Plaid credentials to kind of login. The problem with Perfios and Yodlee today and the approach that they use today is that they ask for your bank username and password. These are now well established companies, they're reputable companies, but even if you ask them, they will admit that it is not the optimal way of doing it because there's always a degree of risk involved when you're talking about handling somebody else's full access credentials just for data puls, right?

Sahil: What it essentially allows you to do is it takes the Plaid approach of essentially having a separate account, so you'll have an account with this account aggregator and you will enter your credentials with that account aggregator and give the consent to this account aggregator to say fetch information from my, let's say HDFC bank account to give it to let's say a lending application so that they have my transaction information so that they can underwrite me, right? and that data flows through an API and it's far more secure way of kind of transferring that data.

Deepak: And then you can revoke that consent-

Sahil: You can revoke that consent, you can manage it. So like I said, it's a consent manager. So these two are two major places where we've made tremendous strides in the last 10 years of building fundamental public infrastructure. That allows for, if you like in the financial system to essentially like an electricity grid, it reduces grid loss, right? It allows you to kind of transfer both data and money more efficiently. And so building on top of that, you can build amazing applications which are yet to, I think, really that Cambrian explosion is yet to happen but the foundations are set and we are on very firm footing.

Deepak: That's a brilliant thing. In fact, I remember in 2011 when IMPS came out and it came out of another initiative of the RBI, which was to link all the ATMs together. So instead of ATMs each individual bank creating links among other banks, RBI set up the base infrastructure at NPCI. And then that became a root of IMPS and IMPS was a little headless because every bank had to implement it on its own. Then UPI came along and out of this layer there's a third parties could develop on top of it. And then now WhatsApp or the IMPS, I mean UPI, which uses the IMPS layer and you could do instant. I remember showing somebody how I could transfer even 10 rupees instantly to somebody else and then they would literally see it in their bank account that very instant.

Deepak: And this for a person from America is almost like magic. There's a statement that said the extremes of technology are indistinguishable from magic. So at some level it sounds like it's truly magical because it's such a leap of faith that he could do this. But I think great points because you've got a base infrastructure at the identity level that helps enable KYC at one layer, it helps enable financial transactions and our financial data to be transferred around. This whole thing I think has now kind of been subsumed into a bunch of smaller things that people now start saying, why does a bank have to bank? So the concept of, I guess, and I want to hear, I mean because I'm also really clueless about this, neo banking. What is it? Is it just branchless banking? Is it contactless banking? Is it banking without cards or with cards? What is neo banking and therefore what kind of scope does it have going forward?

Sahil: Neo banking, it's not really a new concept. I think Neo banking as we understand it today really is technically not a licensed entity and therefore not really a bank. One of the things that I feel is kind of inevitable, if and when these applications were all fairly well funded, do well and gain some user traction, particularly in say the millions of users, I think there's an inevitable day of reckoning where the RBI comes and says, who allowed you to call yourself a bank, because they're technically not licensed entities. Neo banking, as we understand it today, essentially is about building a better user experience layer on top of the same banking products and the same banking infrastructure that any bank has available today.

Sahil: What that means is I'll have some cooler features like add my friends or some social features perhaps and a much better, much slicker user experience. But essentially the underlying products are pretty much the same. In fact, if you look at most neo banks today, the underlying instrument is actually a prepaid account. So just like a prepaid wallet essentially. The neo bank will go and open a large current account with the bank and it'll say, I will run the ledger for this current account and all the sub accounts underneath will essentially roll up to that current account and I will issue prepaid debit cards or prepaid cards against each sub account, right? The benefit of that of course is for up to 10,000 rupees. You need only minimal KYC, you are able to issue a card.

Sahil: And why is this interesting to the bank? I mean, if you really think about it, it's interesting because on prepaid wallets all the float stays with the bank. So essentially the bank is very happy saying all right, you're essentially opening a giant asset for me, it's essentially a zero cost liability and I'm going to essentially be able to lend that money out and make a good amount of money on it, like six, six and a half percent at least, right? And in addition, the interchange on prepaid cards is much higher. It's not capped like debit card interchange and therefore the fee income that is there will be split between say the new bank and the bank itself. So there are two modes of income for them essentially.

Sahil: And on top of that, any other transaction could be charged, etc. Now obviously when you see a new UX, it could be a slick UX for the really elite user per se, or it could be for example, of a vernacular interface for blue collar workers. And different Neo banks are essentially targeting different segments. There's a bank called Yelo, which does the great thing of essentially building a purely Hindi and purely vernacular interface and with audio prompts and things like that. And this kind of UX innovation is something that the bank may or may not be inclined to do, but this company, for example, to go and acquire users will spend time and effort to do that. There are several others who are trying and targeting different segments of the market and that's their prerogative.

Sahil: But really the financial product at the end of it is a prepaid card and a prepaid account. The exception to this of course is Open, which opens a current account for you and Open’s partner is ICICI and they're an SME neo bank, right? And there they're opening current accounts on behalf of ICICI Bank. The role there of the neo bank technically is that of a banking correspondent because they're opening bank accounts and enabling transactions through that application and they can provide, again, really cool features like pay outs, automated payouts or accounts payable, accounts receivable, kind of reconciliation services-

Deepak: Cash management.

Sahil: Cash management, all of that bundled into a nice user interface with a dashboard, etc. But what's really important is that the fundamental instrument is still with the bank.

Deepak: Actually that may not necessarily always be true because if you think about what a bank does, all it does is let's say takes deposits. Now I could put deposit straight into a liquid fund, mutual fund and I can say the liquid fund actually acts like some kind of a deposit account for me. A liquid fund is liquid so I can take money out of it. Currently only banks are allowed to issue cards, but apparently some old archaic rule does exist which allows mutual funds to partner with banks to source an account directly from the mutual fund itself, which allows you as a mutual fund therefore to issue a card against it. A loan can be taken from an NBFC triggered on with a combination of these two elements together. You might not even need the bank because all the bank does is either takes your money, gives you money or helps you pay somebody else. If you did all of these without the bank itself, you'll literally be a shadow bank without actually needing the banking infrastructure for the very same things that you do. But you're seeing neo banking is actually beyond that.

Sahil: There is one part of that where the bank will be involved, which is the mutual fund cannot issue a card without a BIN. A bank identification number and only banks, which are members of those card networks can provide you with a BIN. So the bank will still be involved and because the bank is involved they will have a say in what can and cannot be done. So essentially being able to do all three pieces, totally without the involvement of the bank, is still not possible.

Deepak: And I guess also because the NPCI infrastructure is only open to banks.

Sahil: Yes.

Deepak: So you can't even connect to it if you could.

Sahil: You cannot become a member of the scheme right now.

Deepak: Yes.

Sahil: Yes, and let's not forget that NPCI is co-owned by all the banks. So who's allowed in their control.

Deepak: So I guess it's like a dictatorship that's favorable to the current entities, but even then, what you're actually saying is the whole banking layer has become a pipe and all that the pipe is doing is earning that money on that float, if you may, but at some level every additional function that they're doing, whether it is lending, whether it is payments, whether it is the actual KYC experience of depositing, all the other stuff that the bank was supposed to do on its own can actually be created at a third party level, much more efficient. The bank just provides the base layer and earns money off of it.

Sahil: Theoretically, yes, but again, let's talk about lending, for example. I as a neo bank can't provide credit until I also have an NBFC which is subsidiary and for that, again, the cost of capital will always be a bit of a challenge early on in the life of an NBFC etc. Furthermore, there are lots of things that banks can do that no other entity can do, which is for example, do other based KYC for example, you need a banking license to be able to do that as per law or to be able to … again, participants in, you'll always need a bank sponsor to participate in a new payment network. So Visa, Master, RuPay and UPI are all bank member network only. So that's the second thing. The third thing would be if you want to offer a credit card, that's again a bit tricky because NBFC's are not allowed to issue cards.

Sahil: So you'll have to do a lot of regulatory gymnastics to be able to do that or to circumvent that and we all know how the RBI feels about regulatory gymnastics. So you might be able to get away with it at a low scale but essentially the natural principle of you cannot do indirectly what you're not allowed to do directly will come and bite you.

Deepak: No. In fact, I think I remember this startup in Chennai and what they were doing was very interesting. It was a jeweller, so reduced to gold and then said, we really should digital, so you buy a certain amount of gold, let's say a kg of gold, you pay 30 lakhs hard rupees for it, and then I'll give you a certificate that is a digital certificate that says I own 30 lakhs rupees worth of gold, which is a kg of gold. Now if I want to pay you, I Deepak Shah wants to pay Sahil Kini, I'll say Sahil do you have an account with this person? I will pay you in gold instead of rupees. So I'll pay you a thousand rupees in equivalent gold units if you may. So what would happen is my digital certificate of one kg of gold will come down by a few micrograms perhaps and you now get those micrograms. But those micrograms are not actually physical pieces of gold taken away from my one kg. It is actually the factors that the gold was supposed to be laying in their ward, they will issue a certificate against it and effectively created a payment mechanism from me to you and RBI shut them down because obviously, you can't do this because there's no verification first of all.

Sahil: And this is not a negotiable instrument.

Deepak: It is not a negotiable instrument. I mean, even though I could technically say I'll pay GST on this transfer at some level, but it still would have violated RBA’s perhaps quasi ownership of any transactional currency, which is perhaps why they brought Bitcoin down as well.

Sahil: Precisely.

Deepak: So gold as a negotiable instrument only works in actual physical form, but it won't work in any kind of digital form that's online because our ways are not common.

Sahil: Plus how will the government collect revenue? I mean the tax man cometh always.

Deepak: Yes. And I think that we would argue whether he should cometh or not, but he will cometh.

Sahil: Death and taxes sir, death and taxes.

Deepak: There you go. In fact here's the other thing about technology. It's always been a great leveler, right? So you had the mobile phone and earlier the smaller the phone was the better, now the smarter the phone. And now everybody has smart phones. So you find the lowest economic strata, which was technically not … for them all of this stuff was not available and now they're watching movies on Jio and it's changed, it's leveled the field. But in banking, unfortunately, things have been really lousy for a lot of the people in the lower economic strata. So I've, for example, I had to pay somebody in SBI, he was not in Bangalore. So they said come to the branch and you can deposit this money. I went and stood in line. And then a person came to me and said, “Sir this line is not for you.” So I said, “Is this line for deposits?” And he said, “No, no, yeah, but they are depositing.” I said, “Who’s they?”

Deepak: So there are a lot of these workers, construction workers standing in line to deposit money and I was a little stunned and I said, “They're also depositors like me. So why am I in a separate line?” But there was this they and then I was like, I got a little irritated with that statement and I said, “No, I want to stand in line and see what happens.” So I reached the end of the queue and I said, “I want to deposit so much money in this queue,” and the person ahead of me was having this argument with the teller and she was like, “You don't have a green card.” So he's like, “But I forgot it at home.” And she said, “No you can't deposit money unless you have your green card.” This is a person who has given up his full day's wages to be able to stand in line. They gave only 20 tokens apparently I wasn't told this, they gave only 20 tokens.

Deepak: You stand in line, you get one of these 20 tokens and you're allowed to deposit money into your own account so that your family in your village can then withdraw it from there. And they charge you five rupees for the pleasure, I don't know, for the whole thing. And then I went in there and I didn't have this green card and she didn't ask me for a green card. And she took the money and then said, “Okay, you can go now.” I went back in another time and then I actually had a business idea. But then this thing needs to change because you can't change people. People will only look at a construction worker and say, “I will treat you badly.” We should perhaps change that from a social point of view. But what can Neo banks do to change this?

Sahil: It's a phenomenal question and really it's one of those questions that's endemic, right, to any kind of service business not just banking. But banking it's a little bit worse because, I'm debating whether to say it or not, but it's true banking is very Brahmanical. Most of the banks, top leadership is so and even more so it fundamentally got a class associated with it, right? And whenever you put people in an intermediary kind of physically there, they will want to exercise some kind of power over people who are less powerful than they are. So definitely there are socioe-conomic issues which kind of emerge and most people find it very difficult to even have a conversation around why, for example, if you take a loan and you default on a couple of payments, one of the things that I've heard, because I've talked to a lot of people from that social class about debt and debt restructuring and why they're in debt traps, etc.

Sahil: One of the funny learnings was, “but I paid this month, I was supposed to pay 3,000 to EMI, I paid this month. It was due on 10th I day paid it on 25th and next month I will pay it on time”. But they don't realize that counts as a late payment and there is a penalty associated with that late payment and therefore their amount won't come down or their outstanding won't come down by as much as they would have liked, right? They don't even understand to a certain extent if, unless you can really sit down and explain to them because they're savvy with numbers, they're not necessarily savvy with the structures that we lay on things. And a simple example is flat versus declining rate interest, right? When you tell them 10% they immediately assume flat.

Sahil: Again, the declining interest part is not something that they naturally calculate. So financial kind of awareness around how these products are structured. I hesitate to say financial literacy because I think people who live on the line from let's say a hand to mouth existence are actually financially far more literate when it comes to managing their day to day paisa. But these credit products and insurance products and all the complexity that we kind of layer on is again something that is very hard for them to grasp. So without some assistance, somebody also taught us the difference between flat and declining, right? You weren't born out of our mother's womb understanding this despite the fact that they're both Konkani.

Sahil: The two things therefore that come into picture is one, the experience of essentially taking that middle man or middle woman out and therefore you are directly interacting with the financial instrument itself on the application. And so you can see how much there is and if you want to withdraw, you can withdraw, the button does not discriminate. And if you want to deposit cash, that part will still remain a bit of a challenge I feel but now you see companies like Phone pe essentially trying to do this cash deposit and cash withdrawal through its merchant network. And I really believe that every kirana store in this country and every pan shop can become an ATM and should become an ATM, if you put the right systems in place.

Sahil: So why shouldn't you be able to deposit money into any account if you're able to kind of do these transfers instantaneously now at zero cost, which is inadvisable in my opinion but that's a different topic. But that's one thing, the buttons and an app does not discriminate, right? So that's one thing. The second thing is though most people building those apps are in Bangalore, Bombay, Delhi, and don't belong to this world, so they have a long way to go before they understand what is the right bar at UX? Because the California School of Design will not translate in Kolapur. So to do that well, you need to be able to actually develop a new paradigm of design and that might mean audio first, it might mean video first, it might mean can we create a UPI interface where after I give my fingerprint, I say ‘Deepak ko Rs50 bhej dijie’ and it just goes.

Sahil: Beyond that why are we stuck to the screen, right? So these fundamental questions can be asked and fundamental innovations can be unleashed. And so that's on the UX side of things. The other side is on the product design side itself, on the manufacturing side. And this is something like I said, Neo banks are not in a position to do right now where, let's say there was a liberalization of the licensing regime and more and more institutions were able to kind of, maybe not a full banking license, maybe a small finance bank license and, I feel like the payments bank fiasco was a lost opportunity. But really if you are able to create a new class of financial institutions that can manufacture financial products that are truly sachetized and make sense and are simplified, because what's simple sells quickly. Liquid mutual fund is a great example, right? You're at very low risk, you can't really lose money on an LMF any more than you can lose money on your-

Deepak: You can, technically but-

Sahil: Yes, you can, but technically you can lose money in your bank account as well. And more people have lost money in their bank accounts in the last year than they have in any LMF. So the practice in theory often show two different results. So if you're able to kind of manufacture fundamentally new and simple and well-structured products, and that might mean different things for different people, the financial lives of the literally thousands of Indians, even if you took a cohort of one lakh people, we are looking at 13,000 cohorts in India alone, right? And that's if we believe that our population is only 130 crores, it’s probably much higher. But when you are kind of, the cash flow curve of, this is something that, and pardon me if I'm kind of going on a bit of a tangent, but the financial lives of different people are very different and today the problem is that distribution dictates product design, right?

Sahil: What my distribution costs are because of my brand structure, because of my ATM network, because of my ridiculous amounts that I pay to finical and policies to your co-founder and the ridiculous amount of premium that you would pay to, I don't know, the entire banking vendor ecosystem that is there. The IT budgets are in the thousands of crores. All of these are overheads and the distribution cost is essentially what dictates the product design. And then it's kind of sold mass market off the shelf regardless, right? If you were to flip it and say that from the cash flow curves of the people, I will decide what are the right financial products for each cohort and from that I will decide what is the best distribution mechanism to make the unit economics work.

Sahil: You can make dramatically different types of financial products come to life because the cash flow curve of a textile mill in Surat is very different from the cash flow curve of a school teacher, which is very different from a cash flow curve of a farmer who has only two income spikes in a year, which is very different from a daily wages laborer. So the frequency of income, the amplitude of the income, the spends, all of that has a say in what is the right lending product, loan product or what is the right insurance product because they're exposed to different risks and so on and so forth that you can kind of design and sell. The idea is though, let me ask you a question, right? How many financial products, let's take credit or let's take even investment, how many of them operate on a monthly schedule as a percentage?

Deepak: When you say monthly schedule, do you mean get paid every month?

Sahil: Get paid or pay every month, EMI every month or receive money.

Deepak: I suppose most credit and perhaps not most credit, at least a reasonable point of credit and perhaps investments wise recurring deposits and, so probably 20% of the investment.

Sahil: 20% of investments, but in credit products, how many of them are EMI versus not EMI?

Deepak: I would imagine other than agri products which are 10%, so agri products a cycle, crop cycle, but I'm assuming most of the rest.

Sahil: 90 plus percent of credit products in India are all EMI, you're right. Why? What percentage of our population receives a monthly salary?

Deepak: Possibly not more than 20%.

Sahil: Not more than 10%. So what is the point of an equated monthly installment if you don't have equated monthly income?

Deepak: But we built a software for every month no, so how can we do it?

Sahil: This is the problem and this is exactly why, I mean one of the reasons why we started SETU, right? The whole idea of enabling at least new types of financial products to be built. So these are the two areas where Neo banking can really make a massive difference. Create UX and really Bharat UX and hundreds of UX’s hopefully because how can one app, it's hubris to assume that one application and every bank's application is no different from every other bank's application. How can that application serve the needs of 130 crore Indians where we have 786 different languages where we have 91 different scripts? It's the most diverse country in the world and you're saying one banking app will solve everyone's needs.

Sahil: So that's one and the other is on the design of financial products. How can we have the same type of financial products one size fits all for everybody, which is why you end up with everyone getting blacklisted for credit people not getting formal credit. All of this financial inclusion is a problem because it's designed with the distribution structure in mind, financial products as opposed to the customers' lives in mind.

Deepak: No, that's very true. In fact, I mean, there's so many examples of this. We use a product called NiYO, which is actually a bank currently linked to a Forex transfer application which gives you much cheaper Forex. Now I didn't realize it was actually an underlying bank account on which I could do everything, but what this did was opened out only one aspect of it to me saying, listen, at any point in time I'll tell you what the current exchange rate is, use your card in any currency and I'll give you the best exchange rate possible. That use case, the limiting use case is actually quite beneficial to me when I travel abroad. So if I did that for every possible other application that you can think of or you would actually get a lot more benefit overall than perhaps having a single banking app. Like we were trying to figure out, oh, this particular bank it has this feature, where is this feature? You have to scroll three times then scroll up about twice and then the bottom on the right hand corner there is this little icon that says ‘do this’. And then I was like, oh my God, nobody's going to find this. You can hide a dead body in the second page of Google search results. You could probably have a mass suicide and nobody would discover it if it happened insider underlying layer of a bank app.

Sahil: My favorite joke on this is from yes minster where Bernard essentially says to the Prime minister, I'm going to file this. And the minister says, no, no, no one must ever discover that we did this. And so Bernard says in that case, I'm certain we should file it.

Deepak: Yeah, there you go. So you'll never find something that's just right at the tip of your nose, right? But then I could never understand the reading because I can't see the tip of my nose. But that's probably just me getting old, but going back, yes. So I think these neo banks make a lot of sense and then they grow, earn some money and you've said that that they'll earn money from this massive float that they have. They won't actually earn money from the float, but do banks share any part of the float income that they make as well as apart from, because-

Sahil: As a matter of course, no, but again, look, all of this is negotiation, right? You can find ways to kind of have that conversation with your banking partners if you achieve certain scale or if you have, it's all about leverage. So how you get compensated is a function of how crucial you are as a partner to the bank. So if you're bringing in a lot of money and then you say, no, I'm going to move to another bank, perhaps there is a way for you to kind of negotiate and get more but not as a matter of course. And there's no legal way for banks to essentially share float with you as is. So they'll have to compensate you in other ways.

Deepak: Marketing fees, I guess or something like that.

Sahil: Any number of ways.

Deepak: In fact, I remember PayPal which was effectively taking money in India and earning money off that float at some point was asked to shift toward a wallet ecosystem where the float can only be earned by the bank underlying that. So coming to that I guess this layers on into the next area which I was thinking of, which is there is fraud and there's compliance. And India has typically followed this gating systems. So you say, once you get in, once you do your KYC you're a good man. You can be the worst person in the world, but you have KYC. And you what, they had KYC but the fact is you can get into a system and do all those things and some of that was okay, but the guy who does not have KYC is always going to be the bad guy.

Deepak: This fraud has to move some point I guess. I don't know what use or what do my, I think what this needs is you've got a good detect fraud, you've got a good surveillance, you've got to detect it, you've got to find elements of fraud, find patterns that are frauded and perhaps don't allow the Jamtara's not because people will not do it, but those things can be detected from a technical point of view while they're happening or at least prevent further occurrences from happening especially if you're complaints system in place. And then because fraud always leads to increased compliance requirements, how does this fit into the new banking ecosystem and what happens?

Sahil: I think there's a much larger conversation than just the neo banking ecosystem because fraud is present even in the formal banking ecosystem, there is insurance fraud, there is obviously fraud in the capital markets that puts us all to shame. Therefore, fraud is a much larger conversation. Neo banking is practically non-existent in India it's just starting off and obviously questions will be asked of it because it's new, but the real stuff happens where in the incumbent ecosystem so to speak, right? So the first, I mean I'm no expert on fraud and compliance, but there are certain first principle things that everyone ought to be doing that we're not doing right now. And the first is sharing fraud data, right? Where is the common fraud repository? Because we all know that 90% of the fraud is conducted with the same limited set of individuals.

Sahil: So we should be able to identify at least from, if you're able to all share information saying yes, this guy screwed me over and then he went to the next institution and did the same thing. It's a pattern. Fraud is always a pattern. So if you're able to share information more effectively and you're able to essentially uniquely identify people who are conducting this, then a lot of the fraud can be controlled or mitigated to a large extent, right? That collaboration began to some extent in the credit ecosystem with what is now referred to as the Hunter software that was built by Experian, which is essentially aggregating some use cases of fraud by any customer that is using the Experian credit score. But it's also cute because, well Experian themselves got, yeah, so that's one aspect.

Sahil: And the second aspect is that they don't have that non-market coverage either. So it will need to kind of really extend to all lenders and reporting off that fraud in real time or near real time by the compliance departments again also has to improve. So sharing that information and having a shared fraud database I think will go a long way in getting that started.

Deepak: To be fair there is one and there's a CIBIL suit filed cases data, especially for companies so it's not individual, but you got people who filed a suit for wilful default. So when you identified these wilful defaulters they put out the data. Data is not terribly difficult format to search. But if you did that and you search for a lot of companies and you suddenly find out these are people who have been in this fraud database but have consistently received new loans as well.

Sahil: You only answered your question Deepak. If someone as sharp as you says it is hard to search in that database, imagine what will happen to an assistant risk manager who's paid 45,000 rupees and has to take the train back to Kandivali by six o'clock otherwise it gets a lot worse because the Mira road fast train will not take him back on time. That's his problem and this is the guy who is going to clear the thing, clear the case there. So the data is not enough. The interface with which you can interact with it, can it kind of throw a flag automatically? Is there a fuzzy logic match saying, okay, this closely matches these eight records, but this is the likely one. There's a layer of innovation on top of that data that needs to happen to make it easy to find, to make it easy to surface and to make sure that there are no false positives nor are there any false negatives.

Sahil: That quality measure of statistics to say that how many of those were discovered later is something that doesn't exist right now. No one's throught through this from a true product standpoint to say that, okay, the risk and compliance teams within the banks are my customers. This is what that customer looks like, this is what their day looks like. And so let me design a business that actually helps them do their job better so that the Chief manager doesn't come and take his case.

Deepak: At some point.

Sahil: Three months from now. And same goes for the chief risk officer taking the chief manager's case, which happens when the RBI comes. So that whole chain of screaming is the only incentive right now for them to do their jobs. But if you were to actually make it easy for them to do their jobs, then I'm sure that the fraud rate can actually come down.

Deepak: You're right, compliance and fraud and in these circumstances I guess have, I mean just detection, just the process of doing all this, it's just so much more. It's so different I guess for each bank. But coming to an interesting point that comes across all of this, so as we do more and more surveillance, as we do more and more, and I guess all of us want to collect that data because tomorrow if there is a compliance requirement, we want to be able to say that listen, we have this data and that's what we use to figure out that this person is not a fraud or that I have enough data out to say I know my customer, therefore it's a KYC. But at some layer, this now leads into this whole issue around data privacy. Okay. So the data privacy issue has issue of saying if you've got my data and you are supposed to use it only for X, how do I know that you've not used it for Y?

Deepak: And if there's a data privacy bill that comes in that maybe glorifies this idea too. I mean because bills always come out with initially half-baked ideas and the implementation then makes it even worse. So they might just say, listen, no, you need an OTP before every time you even access this data, which then complicates life both for the guy who wants to share this data and the person who wants that data shared in order to be able to complete a transaction. Because at two o'clock in the morning you're saying ‘OTP kya hai?’. So at some layer data privacy also comes in and sits into this ecosystem and says how does it perform? So what do you think about that space?

Sahil: Yeah, no I think India’s data privacy scenario has been basically the wild West since the inception of the internet till now we haven't really had a data privacy or a data protection bill and really whatever provisions are there in the IT act are woefully inadequate to kind of deal with the kind of stuff that's possible today. And what's really clear is that we need a data privacy law. So there needs to be a law that governs the sharing, collection, use, storage and processing of data. How each of that is done on this data needs to be governed. Now different types of data, different degrees of sensitivity, financial data is sensitive, medical data is even more sensitive, your education data might be slightly less sensitive than your financial data, so there is gradations.

Sahil: So lumping everything together and saying that data privacy is good or data privacy is bad, lacks nuance and the problem is in public discourse, particularly on TV and all of these channels has never kind of focused on the nuance and this is all about the nuance. So how you govern that is kind of key. Now the privacy law that is on the docket and the framework essentially puts user consent or informed user consent at the heart of it saying that the data is the user's property of the user. Anybody who has the data acts as a custodian of that data and therefore is a service provider to the user and must carry out the will of the user should the user say, share this data for this purpose.

Deepak: So in a way the fiduciary of the-

Sahil: It is a fiduciary. They are in fact called data fiduciaries in the law. That is the term. That is the term of art used in the law in the draft bill. So we think that's a definite step forward and it's modeled quite heavily on GDPR and it's our current data protection bill and it has a few extra provisions that some argue is too far, some argue is not enough. What's clear in that law is a lot of power is reposed in an entity that is yet to be formed called the DPA, the Data Protection Authority. And this Data Protection Authority is envisioned to kind of form all the guidelines for data collection, storage, processing, sharing and use. And so the granular details of what a bank can and cannot do, what an NBFC can and cannot do, what a hospital can and cannot do, what a school can and cannot do with your information is yet to be clear.

Sahil: And yes, our track record and regulation has always been to overregulate and be over prescriptive and so the chance of your 2:00 AM sweet nothings being about OTP instead of other more wonderful things is quite possible. Though what is the alternative? Are you saying that we should continue without a data protection regime where any way data is being bought and sold or data cards or USB sticks at a few paisa user record? Because that promiscuous attitude towards data exists today and it's treated like there is no consequence because there isn't. Now we will introduce some consequence and as a result we might introduce some complications and some inconveniences. All regulation can be tuned if you are, because every regulator, at least the RBI and the financial sector regulators have always put the user at the heart of things.

Sahil: And yes, they've also been over regulating it sometime or the other. But what's kind of lost on a lot of the ecosystem is they're also trying to look out for the little guy and the abuse and the fraud that happens at that level is much higher. And so they're trying to protect that and then just like you can't cut only one end of a bell curve, when you kind of limit one end, you automatically limit the other. So when you try and reduce fraud too much, you can also stifle innovation. That's a tight rope walk that every regulator in the world has to do. But I would say what's very, very painfully clear to me is that this is needed now, how it will be operationalized and what needs to be tweaked and how we strike that right balance between accountability and transparency and privacy. I mean, it's a complicated question and I think it would be a stretch to say any country has figured it out, no country has figured it out. And we are the vanguard right now and we will see.

Deepak: Yeah. Well I guess one of the other complications in India is that because of our legal system, we have a lot of delay in enforcement, which is why regulation always trumps enforcement. So you say, listen, it's going to take a poor guy five years to recover money if anything like that happens. So we'll put rules to ensure that it happens. It's a very interesting take on this. But now, as we go forward, I think you've done a lot of this stuff as part of your doing or building up SETU. So tell us a little bit more about what SETU is about and what you're doing in this space.

Sahil: Thank you for that. So the chance to do a shameless plug on your podcast. Basically our theory of change is very simple. If banking and financial services 1.0 was you having to go to the bank or the bank branch or the NBFC branch or the insurance agent in that sense and carry out your transactions. Banking 2.0 was essentially the bank came to your phone or to your laptop or to your computer and you're able to kind of do banking operations on your phone and therefore you're able to kind of do some borrowing, do some investing, do some insurance, etc. But you had to go to the banks application, right? Or to certain distributors that only did distribution of financial services. We believe that banking and financial services 3.0 is essentially where the bank is available in every application, wherever you might need it, right?

Sahil: What that means is if I'm paying off an invoice, I might need credit and therefore loan offers should be readily available at the point of checkout at that time, like it happens in EMI based checkout. You're already beginning to see that where financial institutions are embedding themselves more deeply into the application on Amazon for example, you can get ZestMoney EMI at the point of checkout. In Ola for example, you can see a two rupee ride insurance powered by Acko, and you're already seeing the sachetization and the contextualization of financial products within third party applications. UPI was the first step that allows you to do payments from third party applications from your bank account directly. Here you're able to do now credit and insurance slowly. And these are the proto avatars of what we think banking 3.0 is.

Sahil: But the problem therefore is that on the kind of technology companies side, only the largest technology companies have enough leverage cash and abilities to go ahead and partner with these banking institutions, stitch these partnerships together, spend the time and the developer bandwidth and the effort and the capital to make this partnership come to life and make it successful. And on the other side, essentially the banks themselves only have so much IT bandwidth and therefore will choose only the largest partners to partner with because that's the only partnership that's worth their time. We think that this will go through the same journey that essentially cloud based storage and cloud services went through in the first decade of this millennium where essentially in the year 2000 for you to actually start a technology company, you would have to buy servers, build your racks, build a connection, get a lease line and essentially spend like half a million dollars, before you even get off the ground.

Sahil: So it's half a million dollars in investment, six months of effort and a lot of expertise involved to even get you to a point where you're going to get started doing a transaction. But AWS came and destroyed all of that. It made it basically practically free to get started and all you have to do is have a credit card on file and you could spin up a server anywhere and get going. The same what AWS did for cloud storage and storage and servers, storage and computing, we believe that SETU can do for all financial services. So essentially it's like BWS, it’s ‘banking web services’ if you want to kind of keep it more simple. So the idea is for you to provide essentially what is banking? It's three things, right? It's account services. So being able to open and manage any kind of account. So that could be a loan account, a savings account, a current account, an insurance policy, a mutual fund folio, all of these, even a prepaid wallet, even a crypto token or gold or any of these are accounts in our books that has a ledger, right, that's essentially associated with it. The second is payments. Payments is money in, money out and money in and money out essentially has three layers. Addressing payment address, and we talked a little bit earlier in this about how we address is different from an account infrastructure, addressing authorization, sorry, authentication and authorization. So all three layers being able to do it across multiple schemes, any NEFT, IMPS, RTGS, Visa, Master, Rupay and UPI. And the third would be data.

Sahil: So data would be in the case of this, the account aggregator is something that we've spoken about again. Can I provide financial information so that underwriting for credit or underwriting for insurance can happen seamlessly and immediately. So account infrastructure, payment infrastructure and data infrastructure, all laws APIs, so that any company wanting to offer financial services and we are hosting essentially the banks. The banks are like the suppliers of these APIs and we essentially make sure that the APIs that they give are consumable. Because today if you go and ask for a bank for API, they'll first tell you open a current account, open an account, and then I'll give you some APIs. That itself will take you like a couple of months.

Sahil: And then after that, the APIs that come to you come to you on a PDF file and no one tells you what API you should use if you want to build this, right? And there's no developer tooling. So what we do is we work with the banks, we host all these APIs and then we make sure that those APIs are in a specification that is standardized and consumable and we create wonderful, delightful documentation around it. Sample apps, standardized error codes, SDKs and dedicated developer support so that you really get the red carpet treatment whenever you're integrating with banking and financial services APIs. So we want to take it from a year long process that can be like pulling teeth and you have to really navigate multiple teams within the bank.

If you're going to launch a product in partnership with the bank you have to work with at least five teams within the bank; the business team, the risk team, the compliance team, the legal team, the IT team and the IT vendor. So the whole aspect is painful for financial products and banks are not used to building technology for the consumption by third party applications because who builds technology at banks? If you really think about it, it's the banks’ vendors, not the banks’ employees, right? And who pays the banks’ vendors? The bank. So therefore for the vendor, the customer is the bank and they will always build software so that the bank can use it, not so that a third party application can use it.

Sahil: We solely exist to be able to play that role of bridging the financial institutions and any third party application that wants to provide this financial infrastructure be it loan, be it credit, be it investments in children's or even classic CASA accounts. We've started with the bill payments offering. We'll soon be launching accounts offerings and soon after that data offerings as well. And we really see this as a long-term 30 year journey of building the connecting infrastructure or the bridges between financial institutions and other applications. And that's pretty much why we called ourselves SETU because SETU Sanskrit word for bridge and that's what we do.

Deepak: Wow. This is pretty interesting. So effectively you're like the, instead of calling it a payment gateway or a banking gateway or a gateway to the large financial institution that has regulatorily really protected.

Sahil: Correct.

Deepak: So it may be the insurance company, it may be a bank, it may be a mutual fund, but they need to interact with the little guy and you're sitting there in the middle saying, I want to standardize your experience and will even give you a sandbox to help you figure these things out by yourself and only when you're ready you go start making your negotiations eventually with the bank.

Sahil: Precisely.

Deepak: Pretty interesting, I like this. This is a very interesting SETU and I hope, I wish you all the very best Sahil, it’s been a phenomenal conversation and I think we'll obviously have a lot more so when Capitalmind becomes a bank, a mutual fund, an insurance company.

Sahil: Amen to that.

Deepak: Yes. I think this has been a fun conversation. Thanks all of you for listening in. Thanks a lot Sahil for coming in.

Sahil: Thank you. Thank you for having me.

Deepak: Thanks. Great. So we are all available on Twitter. You can find both of us. I think Sahil is SahilKini on Twitter. I am Deepakshenoy. We are at Capitalmind_in and throw us your bouquets and leave the brick bats for another day, but the brick bats are also welcome. Please get in touch and thanks so much for listening.