AI agents for operations leaders in retail banking in India

Periodic KYC updates, inoperative account reactivation, deceased claimant settlements and multilingual complaint handling are where AI agents help Indian retail banking operations leaders. Agents check documents, trace records and draft in the customer's language; people decide claims, freezes, credit and complaint outcomes. The RBI's rules and the Digital Personal Data Protection Act 2023 apply.

What fills an Indian retail banking operations leader's week?

KYC updates. The RBI requires banks to update customer due diligence periodically, more often for higher-risk customers. Very large numbers of accounts come due on a rolling basis, and customers who do not respond face restrictions. Operations chases, checks and records each update across branches, digital channels and business correspondents.

Inoperative accounts and unclaimed deposits. Accounts without customer-initiated transactions become inoperative, and long-unclaimed balances move to a fund maintained by the RBI. Banks run reactivation drives and trace the owners or heirs of old deposits.

Deceased claimant settlements. When an account holder dies, the bank settles with the nominee, the survivor on a joint account or the legal heirs, under RBI expectations on simplified procedures and turnaround times. Families often arrive with incomplete documents and in distress.

Scale and language. Basic accounts opened for financial inclusion mean many customers are first-time users of formal banking, and they write and speak in many languages.

Where can AI agents help Indian retail banking teams?

• KYC update processing (depth 2 and 3). The agent checks submitted documents against the bank's requirements and the customer's risk category. Simple self-declarations of no change are recorded at depth 3 within set limits; anything else is drafted for a person.
• Inoperative account reactivation (depth 2). The agent identifies the account, checks the identity documents received and drafts the reactivation for a person to approve, because reactivation is a known route for fraud.
• Deceased claimant file preparation (depth 2). The agent lists the deceased's accounts, lockers and deposits, identifies the nominee or survivor, checks documents against the bank's settlement checklist and drafts the request for anything missing.
• Complaint classification and drafting (depth 2). The agent classifies complaints, sets the deadline, gathers the account history and drafts the reply in the customer's language.
• Unclaimed deposit tracing (depth 1). The agent searches records for the owners or heirs of long-unclaimed deposits and drafts outreach for a person to approve.

Inoperative account reactivation shows where caution belongs. The work is document checking, which suits an agent. But fraudsters target dormant balances, so reactivation stays at depth 2 even when the documents look perfect.

What stays with people, and why?

Claim settlements to nominees and heirs stay with people. Paying the wrong person cannot easily be undone and can lead to litigation.

Account freezes and restrictions for non-compliance with KYC stay with people. They cut customers off from their money.

Credit decisions stay with people, under the bank's credit policy.

Complaint outcomes, including any compensation, stay with people. Customers can escalate to the RBI's ombudsman.

Approving new versions of the agent stays with people. Because the agent works in several languages, an update that improves one can weaken another. Our answer on how to compare two versions of an AI agent explains replaying each language's cases before release.

Which rules apply in India?

The RBI regulates banks, including customer due diligence and periodic KYC updates, inoperative accounts and unclaimed deposits, deceased claimant settlement, customer service, grievance redress through its ombudsman, outsourcing and IT risk. Its outsourcing and IT governance expectations cover an agent and its supplier. The RBI also requires certain payment data to be stored in India, which affects where an agent may process transaction data. SEBI and IRDAI matter where a bank distributes investment or insurance products.

The Digital Personal Data Protection Act 2023 governs personal data, with duties on notice, consent or other permitted grounds, purpose limitation, accuracy, security safeguards and breach reporting, and rights for individuals to access and correct their data. KYC files and claim files contain identity documents and family details, so access should be narrow. Our guide to the DPDP Act and AI agents in India covers the detail.

Money laundering controls follow the FATF Recommendations. India has no single AI law.

This is an explanation of the rules, not legal advice; confirm how they apply to you with counsel.

How should you start in the first 90 days?

Weeks 1 to 4: start KYC updates in shadow on one region's due list. Score the agent's document checks against the decisions branch staff made.

Weeks 5 to 8: certify no-change self-declarations at depth 3, limited to accounts below a balance threshold you set in rupees. Begin deceased claimant file preparation in shadow.

Weeks 9 to 13: add complaint drafting in the two languages that carry most volume, certified separately. Keep reactivations at depth 2. Follow deploying an AI agent: train, shadow, certify, widen. For a market with similar inclusion pressures, see our brief on operations agents in Nigerian retail banking.

Where Pryme fits

In Pryme, the core banking record arrives as a sealed source, a death certificate verified through an official channel as signed, and a claimant's statement about family relationships as declared. The claim agent shows which class supports each step. Settlement sits above its depth, so it is held for a named officer, and every verdict is filed.