AI Report: 10-20% Branch Time Lost to Routine Work in Indian Banks

For investors and bank strategists, the August 15 FICCI-IBA-BCG report quantifies an inefficiency overhang: 10-20% of branch time is lost to routine servicing and a single loan journey can require 50+ manual fields. That translates into a measurable cost and throughput opportunity across public and private lenders as agentic AI moves from pilot to production.

On 15 August 2026, the Federation of Indian Chambers of Commerce & Industry (FICCI), the Indian Banks' Association (IBA) and Boston Consulting Group (BCG) released a joint report arguing that Indian banks must move beyond digitising structured processes and fundamentally redesign their operating models to unlock the next wave of productivity from artificial intelligence. The central claim is that maximum productivity gains from AI will come not from simply automating existing workflows but from rewiring bank operations, so that repetitive tasks and activities with predictable outcomes are automated while human resources are redeployed towards decisions involving greater complexity, judgement and value.

“

For investors and bank strategists, the August 15 FICCI-IBA-BCG report quantifies an inefficiency overhang: 10-20% of branch time is lost to routine servicing and a single loan journey can require 50+ manual fields.

Despite significant investments in digital infrastructure over the past decade, banks continue to rely on substantial manual intervention, physical documentation and complex processes. The report quantifies the drag: a single loan journey can require more than 25 documents and 50 or more manual data-entry fields, while six to eight layers of manual review may be required. Banks also dedicate thousands of customer service agents to repetitive interactions, and 10-20% of branch time can be consumed by routine servicing requests. These figures make the case that digitisation alone has not eliminated operational friction, only digitised parts of it.

The report distinguishes between two eras. The first phase of banking digitisation focused on "coding the known": automating finite, rule-based journeys involving structured information. That phase delivered significant gains in areas such as digital payments and identity verification, but conventional automation remains constrained by the rules and permutations that have been explicitly programmed. The emerging AI era, by contrast, will be driven by agentic execution, where AI agents process unstructured information such as voice, documents, images and free-form text and execute tasks across multiple permutations based on intent rather than predetermined rules. This is a step change from robotic process automation to systems that can interpret and act in less scripted environments.

The report also identifies conversational banking as the next frontier. As customer expectations shift, banks that have built digital rails but still route simple requests through human channels will face pressure to deploy conversational agents that can resolve queries, initiate services and complete tasks without predefined scripts. The implication is not merely cost reduction; it is a redistribution of human attention toward complex customer needs, risk judgement and relationship management.