In December 2024, the Reserve Bank of India set up an eight-member committee to work out a framework for AI in banking, NBFCs and fintech — chaired by Prof. Pushpak Bhattacharyya of IIT Bombay, with members drawn from RBI itself, IIT Madras, the Ministry of Electronics and IT, Trilegal, HDFC Bank and Microsoft India. The committee's report, titled the Framework for Responsible and Ethical Enablement of AI (FREE-AI), was published on 13 August 2025.
The seven sutras, again
FREE-AI is built on the same "seven sutras" as India's national AI Governance Guidelines, published the same year: trust, people-first governance, innovation over restraint, fairness and equity, accountability, understandability by design, and safety and resilience (see our companion post on the national guidelines). That's not a coincidence — RBI's sector framework is explicitly built to sit underneath the national one, not compete with it. What FREE-AI adds is a financial-sector-specific structure: six pillars (Infrastructure, Policy and Capacity for enabling AI adoption; Governance, Protection and Assurance for managing its risk) and 26 concrete recommendations mapped across them.
The one recommendation with real teeth
Most of the 26 recommendations read as exactly that — recommendations to RBI, not law until RBI formally notifies them as regulation, so treat the committee report itself as a roadmap rather than a rulebook you're already bound by. The clear exception, flagged in legal commentary on the report, is stated in unambiguous terms: every RBI-regulated entity — bank, NBFC or fintech — is expected to adopt an AI governance policy approved by its own Board before it deploys AI at any scale. If you sell an AI tool into an Indian bank or NBFC, that Board-approval requirement is the one line in this report your buyer's procurement team will actually ask you about.
Why this lands on customer-support AI specifically
The committee's own survey of the sector found customer support to be one of the most common AI use cases among regulated entities already — and flagged that most of it is still small-scale pilots, often an internal chatbot rather than anything customer-facing at scale. That matches what we've found building out this site's own voice AI coverage: India-specific voice-agent vendors for regulated call centres are still a short list (see our voice AI & call-centre agents shortlist) next to how far WhatsApp and chat automation have already gone. FREE-AI is effectively RBI getting ahead of that gap before it widens — a Board-approved AI policy is easiest to write before your contact-centre is running an AI voice agent, not after.
What we don't know yet
RBI hasn't published an implementation timeline or a compliance deadline as part of the committee report itself — the report is the committee's recommendation to RBI, and it's now on RBI to decide which of the 26 to formally notify, and when. If you're evaluating an AI vendor for a regulated financial business today, ask them directly whether they have a Board-approved AI governance policy on file rather than assuming FREE-AI already forces the question — check RBI's own circulars for what's actually been notified before treating any of this as a settled compliance requirement.