Banking AI race moves from chatbots to deep institutional layers

Artificial intelligence in banking will shift from customer-facing chatbots to deeper operational layers, treasury, risk and compliance, where the biggest and most durable gains are expected, Livemint reports. At the ET BFSI CXO Conclave 2026, executives said AI is becoming an operating and customer-engagement layer, moving beyond isolated automation to integrated platforms.

Real-time personalisation, fraud detection and credit underwriting are key use cases. JPMorgan has doubled transaction-screening volume while halving manual checks. In India, shared infrastructure through Aadhaar, UPI and Account Aggregator creates a foundation, but institution-specific AI work remains critical for treasury and compliance, Livemint notes.

Executives at the conclave cautioned that data quality, privacy, explainability and governance remain challenges. Boards are demanding clearer returns, with large-scale AI investments expected to deliver full results in three to five years. Human oversight remains essential for complex decisions in lending, wealth management and insurance.