How BankBazaar serves 70 million users: AI credit Literacy and Gold Secured Loans

An interview with Adhil Shetty by Sean Murphy

Every month, BankBazaar generates over a million videos.Each one is different, addressed to a single registered user,delivered in one of twenty Indian languages, walking thatperson through the specific reasons their credit score hasmoved and what to do about it. The AI reads the user’s creditreport, identifies the levers, and produces an explainer tohelp them improve their credit score. One user in a smallertown might be told that their score has dropped twentypoints because they have maxed out one of their creditcards, and that redistributing their spending could lift thescore by fifteen points over three months. Another user mightbe guided through the precise documentation needed todispute an alleged late-payment on a mortgage. The videosgo out across the country, into over 20,000 pin codes, andthe production line behind them is software.

Adhil Shetty, who co-founded BankBazaar in 2008 and isthe current chief executive, calls this kind of product “AInative”. He means something more specific than thetypical usage of the term. AI native, in Shetty’s framing,does not describe a company that adopted AI early, orone built on top of large language models. It means aproduct that could not exist without AI at all. Producing amillion tailored videos a month is inconceivable withoutAI.

Personalised credit coaching at this scale, in this manylanguages, is not a digitised version of something ahuman team could have done more slowly. It is acategory that opened up only when the underlyingcomputation became cheap and capable enough toserve a user base the company has spent close to twodecades building.

BankBazaar has more than 70 million registered users, a customer base larger than the population of France. The company began in 2008 as a marketplace, comparing rates from Indian banks and earning a commission when consumers signed up for credit cards or loans. Its first pivot, several years in, was to build co-branded cards with partner banks, where BankBazaar’s own brand sat on the plastic and revenue came not from a one-time referral fee but from an interchange-style annuity on consumer usage. Sequoia, now Peak XV, was an early backer. Amazon, Eight Roads and Experian followed.

The new subscription product marks a third act. Consumers below the 700 mark on their credit score, the segment most willing to pay for help climbing into prime territory, sign up for a monthly fee. The product analyses their file, generates the personalised video, and tracks their improvement against modelled outcomes. Prime and super-prime users mostly do not buy it, Shetty notes, because they believe, often correctly, that they already know what they are doing. The buyers are people on the way up, and they are not necessarily concentrated in Mumbai, Delhi and Bangalore. The product moves through the country in roughly the same distribution as awareness of credit scores itself, which has spread to smaller towns and non-urban areas as advertising for car loans and mortgages has saturated the market with the idea that a 750 score unlocks a better rate.

Their new subscription product is driven by a behavioural shift that BankBazaar has identified. Around 200 million Indians, the top fifth of the consumer market, are now willing to pay subscriptions for financial advisory tools. That was not true two years ago. People pay for entertainment, for e-commerce, for video apps. The idea that they would pay for help managing their own credit score was a hypothesis. Now it is an observable and growing revenue line.

Alongside the AI video product, BankBazaar’s other recent launch is more of a blast from the past. The company is now offering gold loans, probably the oldest form of secured consumer lending in India. Indian households hold an estimated 25,000 tonnes of gold, one of the largest concentrations in the world, mostly in jewellery acquired over generations as both ornament and savings. The pawnbroker who lends against gold has existed in many neighbourhoods for centuries and, in the unregulated form that still takes the bulk of the market, often charges three or four per cent interest a month, structured opaquely enough that the customer does not always realise what they are paying.

BankBazaar’s regulated version looks quite different. Tech-assisted purity assessment now takes around fifteen minutes. Loan-to-value typically runs around 70 per cent. The Reserve Bank of India sets the reserve price at 90 per cent of the prevailing gold rate, so the lender cannot quietly sell a customer’s jewellery for less than it is worth. A series of formal notices precedes any auction. And for the consumer sitting in the 650 to 750 credit-score band, where unsecured personal loans are hard to access, the gold loan becomes a cheaper option.

It is, in its own way, the same idea as the AI video. An old asset, an old behaviour, and an old form of security, organised through regulated infrastructure and digital delivery so that they work for the customer rather than against them. Shetty is clear that AI is now moving through every function in the company, from engineering, where smaller teams produce more, to the call centre, where virtual agents handle volumes a human team could not. His description of the internal process is procedural. Each team examines its tasks, identifies where AI can step in, and builds workflows so the next time the same job arises, the AI handles it. Tool access matters, he says, but the harder work is department heads sitting with their teams to consciously design AI-first systems.

Close to two decades in, BankBazaar’s third act looks promising. The marketplace model is mature and the co-branded card business is mature, and their plans for growth are forward thinking and driven by what their customers actually need. BankBazaar's approach is to layer products onto a user base of seventy million that already trusts it with their financial data, and to use AI to make those products economically viable at a scale and granularity no human team could ever replicate. 

Eighteen years of work have given BankBazaar something most fintechs do not have: a seventy-million-strong customer base whose needs now point the way forward. The company's job from here, is to keep listening and evolving.