For years, getting a loan meant hunting for bank statements, scanning PDFs and waiting days for an answer. Increasingly, lenders are replacing that paperwork with your consent to access your financial data.
The shift is being powered by the Account Aggregator (AA) framework, which lets borrowers securely share financial information with lenders without the AA storing it. The payoff is faster decisions—and, in some cases, better loan pricing.
Richer data is also helping AI-driven underwriting models distinguish disciplined borrowers from risky ones. As of December 2025, 64 lenders had onboarded onto the Reserve Bank of India's (RBI) Unified Lending Interface, a centralised gateway connecting lenders to AAs and other data sources. They were using more than 136 data services to assess borrowers beyond a traditional credit score, according to a Press Information Bureau backgrounder dated 13 May.
Mint spoke to industry players to understand how sharing AA data can work in your favour—and what RBI and Sahamati proposals are trying to fix.
Sharing data can get you a better deal
According to Sahamati, the RBI-recognised self-regulatory organization for India's AA ecosystem, the network facilitated about ₹1.47 trillion of lending across 15 million loans between April and September 2025. Nearly one in 10 personal loans now goes through AA-enabled data sharing.
Lenders can assess cash flows, income stability, spending patterns and existing liabilities in real time, giving them a fuller picture of a borrower's ability to repay than a credit score or uploaded PDF, said Sachin Seth, regional managing director of CRIF India and South Asia.
They can also track when and how consistently a salary lands, examine mutual fund holdings and calculate debt-to-income ratios.
For self-employed borrowers, people with irregular income or those new to credit, that richer picture can mean higher approval odds, Seth said. “In some cases, stronger evidence of financial health can also translate into better pricing, higher credit limits, or more tailored credit offers.”
But better data is not a guaranteed better deal.
Kapil Garg, managing director, Mufin Green Finance, said AA is only an enabler of better decisions, not a guarantee. Pricing and approval still depend on a borrower's credit profile, repayment capacity and each lender's framework, among other factors.
What are you consenting to?
Consent is triggered each time you share data for a specific purpose, such as a loan application. Only the lender you consent to can access it; a different lender—or the same lender for a different purpose—needs fresh consent.
The data shared is limited to what you have linked and approved on the AA app, across categories such as bank accounts, mutual funds and insurance policies. Linking is a one-time setup, but the transfer happens each time you approve it.
The AA transfers the data without storing it. Once it reaches the lender, however, you generally cannot see how it is processed or which third parties handle it.
That is where the system's biggest blind spot may lie.
Yatin Pednekar, co-founder and chief of products at Mobicule Technologies, a debt collection platform, said, “A real safeguard missing today is purpose limitation, meaning that data collected to assess your loan shouldn't quietly get reused for debt collection or marketing later, unless you separately consent to that specific new use.”
While lenders are legally forbidden from sharing your data for these purposes, in practice it is difficult for an outsider to monitor.
Rajat Deshpande, chief executive and co-founder of FinBox, a credit infrastructure company, said customers grant consent because the process is convenient, not necessarily because they have understood everything that happens to their data afterward.
What is being proposed
The RBI released draft rules last week aimed at reducing hidden risks in AI-based lending. The rules would require lenders to explain automated decisions such as loan approvals, disclose when customers are interacting with AI and hold bank boards legally accountable if something goes wrong.
Sahamati proposed its own framework in June, focusing on what happens to a borrower's data before it reaches a lender's AI model.
Kiran Gopinath, chief innovation officer and head of Sahamati Labs, said, “The framework introduces guardrails so consumers can track how their data is processed. Agents can run amok if you don't put guardrails in place. So, we're able to track the processing of the data as well.”
Under the proposal, a borrower could connect bank accounts through AA as usual. But the data would first pass through what Gopinath calls a "confidential computing" environment—a secure, neutral space where the AI does its work.
“You would know every action that the agent has performed, and that action produces a processing receipt,” he said. “So if the lender is looking at income assessment, you would know that a particular action has been taken, for example they've assessed your income.”
Why loan rejections feel like a black box
Both frameworks are also trying to tackle opaque loan rejections.
Pravin Jadhav, founder of unicorn stockbroking app Dhan, recently said on X that a leading private bank rejected his home loan application after classifying him as a ‘high risk borrower’. He said his credit score was more than 800 and that he may be in the top 0.1% of the bank's customers by assets.
It is not confirmed that AI made that particular call. But the episode highlights the kind of unexplained rejection both frameworks are trying to prevent.
Under the proposed Sahamati framework, Gopinath said, “He would know what data was processed and what processing took place on the data he shared via AA – whether it was for income assessment, debt assessment etc. While the decision might be taken using other sources of data like credit bureaus, and the lenders might not share the full picture, the framework empowers him to know what was processed from his AA data.”
If a borrower disputes a rejection, a grievance officer could trace that trail instead of the bank having to reconstruct what happened after the fact from its own internal logs, he added.
Today, under Section 11 of India's DPDP Act, a borrower can request a summary of their data and how it has been processed, but not the actual logic behind an AI's decision, said Anshul Verma, partner at SKV Law Offices.
Unlike the European Union, India does not currently give borrowers a legal "right to explanation" for automated decisions. The RBI's explainability push is a draft guidance for now, not an enforceable right, he added.
Verma also flagged another risk: if too many lenders rely on the same handful of AI models or scoring vendors, a flaw in one of those models could become a systemic problem.
What can you do now?
Until these frameworks are finalized, borrowers should stick to RBI-recognized lending apps, Verma said. Protections such as data localization and limits on apps accessing contacts or photo galleries apply only within that regulated space.
Once a loan has been repaid, borrowers can ask the lender to delete their data. This is already a right under India's data protection law, though it applies once the lender has no further legal reason, such as RBI record-keeping rules, to retain it, Verma added.
Deleting consent on an AA app only stops the sharing of fresh data. For information already with the lender, borrowers can submit an erasure request.
Deshpande advises borrowers to check what data they are being asked for and for how long, and to avoid broad or recurring access when one-time consent would do the job.
AA is making financial data more useful to lenders—and helping borrowers get loans faster and potentially at better prices. Until the proposed safeguards are finalized, be deliberate about what you consent to.