Quick Read
Summary is AI Generated. Newsroom Reviewed
-
Rajesh Kumar was denied a loan by an AI lending app without explanation despite good credit history
-
India’s AI lending market is growing rapidly, projected to reach $9.5 billion in 2024
-
RBI mandates lenders to explain credit decisions and provide data usage clarity from May 2025
At 2:47 pm on a Tuesday, Rajesh Kumar's expansion plans hit a wall. The electronics shop owner from Pune had saved for two years to capitalize on India's smartphone boom. With the iPhone 16 launch driving repair demand, he needed Rs 2,00,000 for advanced diagnostic equipment and component inventory.
His tech-savvy nephew suggested skipping the bank queue entirely. "Uncle, these AI lending apps give decisions in minutes." So Rajesh downloaded the app, uploaded his documents, and waited.
Three minutes later: "Application rejected."
No explanation. No human contact. Just an algorithm that somehow decided this entrepreneur with perfect payment history wasn't worthy of Rs 2,00,000.
This paradox defines India's AI lending revolution: technology promising financial inclusion while simultaneously excluding people without explaining why.
AI Lending In India: Power, Scale And Trust Gap
Rajesh's experience reflects a broader transformation reshaping Indian finance. Machine-learning algorithms now process millions of credit applications daily, analyzing everything from UPI transaction patterns to mobile recharge timing. These systems have discovered, for example, that people who pay electricity bills on specific dates or shop online for specific types of luxury goods are statistically better borrowers.
The numbers tell a remarkable story. India's alternative lending market is growing 26.3% annually to reach $9.5 billion in 2024. The digital lending platform market expanded from $400 million in 2023 to $2.38 billion by 2030 (Research and Markets, 2024; Grand View Research, 2025).
The advantages are undeniable. Instant decisions replace week-long bank processes. Massive scale enables serving millions simultaneously. Alternative data sources reach previously excluded segments. India now boasts thousands of online lending NBFCs, making this possible.
But here's the catch: these AI systems have become so sophisticated that even their creators can't always explain individual decisions. When algorithms discover complex patterns linking grocery shopping timing to credit default probability, human intuition hits its limits.
This opacity creates real problems. Rajesh represents millions of creditworthy borrowers getting rejected without understanding why. In a sector built on trust, algorithmic mysticism becomes a liability.
Why Regulators, Borrowers Demanding Clarity
Regulators are finally responding to stories like Rajesh's. The Reserve Bank of India's Digital Lending Directions, issued in May 2025, essentially tell lenders: explain your rationale or face consequences. The rules mandate Key Fact Statements before loan execution and clear communication about data usage (RBI, 2025).
Governor Sanjay Malhotra has been explicit: innovation cannot come at the expense of consumer protection. In fact, the central bank is creating a public repository of authorised lending apps specifically to combat algorithmic opacity.
International precedents offer sobering lessons. Wells Fargo faced devastating legal consequences in 2022 when their algorithm discriminated against qualified borrowers. The bank couldn't explain their AI's decision-making process, making legal defense nearly impossible. Lawsuits, regulatory penalties, and reputational damage followed (Robert F. Kennedy Human Rights, 2025).
Research reveals similar bias patterns globally. Facial recognition systems show reduced accuracy for darker skin tones. Hiring algorithms demonstrate systematic bias against women and minorities. When training data reflects historical prejudices, AI amplifies rather than eliminates discrimination (Legal Developments, 2024).
How Explainable AI Builds Trust
Six months later, Rajesh encountered a different kind of lender. When he applied, instead of a cryptic rejection, he received detailed feedback: "Your debt-to-income ratio of 47% exceeds our 40% threshold. Reducing monthly obligations by Rs 6,000 would qualify you for approval. Your consistent payment history and growing business revenue work in your favour."
For the first time, Rajesh understood exactly what needed to change. Two months later, after restructuring some personal loans, he reapplied and got approved.
This is explainable AI in action. It's not about dumbing down algorithms but making them show their reasoning. Feature importance analysis reveals which factors matter most. Counterfactual explanations specify exactly what changes would alter outcomes. Explainable models create human-readable versions of complex decision trees.
Effective implementation requires three foundations:
Auditability ensures every decision traces back to specific data points. No black magic, no secret formulas. When someone gets rejected, both the borrower and regulator can see exactly why.
Informed consent means borrowers understand what data is being used and how. The RBI's new framework requires explicit permission before accessing device features and enables consent revocation.
Feedback mechanisms provide clear paths for error correction. When AI makes mistakes, and it will — borrowers need straightforward processes to challenge decisions and fix incorrect data.
Innovating Responsibly: Winning In Regulated AI Future
Winning lenders aren't fighting transparency requirements. They're embracing them as a competitive advantage.
Several major Indian banks, NBFCs, and MFIs are already adapting. Several institutions have invested heavily in explainable AI for credit decisions, providing detailed reasoning for both approvals and rejections. These early movers are building customer trust that translates into long-term relationships.
This transformation requires fundamental changes. Development teams must prioritize fairness alongside accuracy. Compliance teams need enough technical literacy to question model outputs meaningfully. Risk management frameworks must monitor for bias and model drift in real-time.
When customers understand lending decisions, they develop stronger institutional relationships. Transparent AI builds trust. Trust drives loyalty. Loyalty generates sustainable growth and profitability.
The True Test of India's AI Lending Revolution
Today, Rajesh runs a thriving electronics repair business. He got his loan, expanded operations, and now services the latest smartphones and laptops. His success began when a lender could explain why he didn’t qualify for a loan instead of just saying no.
Trust and transparency will determine winners in India's AI lending revolution. The most successful institutions won't see AI as merely an efficiency tool but as a responsibility to create equitable credit access.
As algorithms make increasingly consequential financial decisions, explaining credit decisions becomes fundamental to sustainable business models. In a market where AI might determine whether someone can expand their business or buy a home, showing the "why" behind decisions isn't optional.
India's digital lending transformation is remarkable. But its true measure won't be transaction speed or processing scale. It will be whether an electronics shop owner in Pune can understand why he deserves a chance to grow his business in India's booming tech economy.
That's when we'll know we've built something worth celebrating.
The article is authored by Joydip Gupta, APAC head, Scienaptic; and Chandan Pal, CMO, Scienaptic.
Disclaimer: The views expressed here are those of the authors and do not necessarily represent the views of NDTV Profit or its editorial team.
RECOMMENDED FOR YOU
US De-Minimis Rollback Can Raise Landed Costs By Up To 15%, Cut Indian Shipments By 25%
Aug 21, 2025
Coolie Box Office Collection: Rajinikanth’s Film Set To Cross Rs 200 Crore In India
Aug 18, 2025
50 Years Of Sholay: How Much Would The Megahit Make Today And More
Aug 15, 2025
Money Wise: Rs 1 Crore Is Now A Pitstop
Aug 03, 2025
POPULAR