Fortune India Banking & NBFCs Summit 2026: AI is reshaping financial services beyond automation, say industry leaders

Artificial intelligence (AI) is moving beyond assisting employees with routine tasks to independently handling financial processes and customer interactions, prompting banks and non-banking financial companies (NBFCs) to rethink how they operate, industry leaders said at the Fortune India Banking & NBFCs Summit 2026.

During a panel discussion titled Reimagining Finance in the AI Era, Pratik Shah, Partner and National Leader for Financial Services at EY India; Pankaj Gadgil, MD and CEO of Aditya Birla Housing Finance Ltd and Head of Digital Platforms and Payments Strategy at Aditya Birla Capital Ltd; and Aman Goel, Co-founder and CEO of GreyLabs AI, discussed AI’s growing role in financial services.

Gadgil said AI was improving customer experience, reducing costs, creating operational capacity and helping financial institutions optimise capital utilisation.

“The opportunity here is completely reshaping the way we think about the sector,” he said, highlighting the shift from AI-assisted operations towards autonomous agents.

However, he stressed that this transition would require appropriate guardrails, governance and monitoring to ensure AI systems operate within clearly defined boundaries.

From AI assistance to autonomous processes

Goel said businesses were increasingly redesigning processes around AI rather than simply using it to help employees perform existing tasks.

“Can this process be redesigned for AI rather than AI being a tool to assist the human?” he said.

Citing GreyLabs AI as an example, Goel said the company, which has around 125 employees, uses AI to run its finance function, including processes involving financial data and book closure.

He said systems could integrate sales proposals, contractual terms and customer consumption data to automate calculations and identify configuration errors, reducing manual intervention.

The shift is also evident in financial services, spanning telesales, customer support, collections and insurance renewals. Companies are moving beyond using AI agents to qualify leads towards enabling them to guide customers through processes such as opening accounts, Goel said.

He added that understanding Indian consumers would be critical to adoption, particularly as voice AI expands across languages and geographies.

Data, governance and trust

Shah said AI had evolved beyond isolated applications to become an enterprise-wide capability, with financial services companies increasingly focused on scaling deployments.

“AI has now moved beyond use cases to an enterprise-wide capability,” he said.

Shah emphasised that trust and governance must be incorporated into AI systems from the outset. He also advocated a phased approach to autonomy, starting with simple, repetitive tasks before extending AI's responsibilities to more complex activities.

Gadgil said data readiness was equally important, as AI systems need access to relevant information to understand customer requirements and provide personalised responses.

“The preparedness for the organisation, both from the standpoint of the platform itself and the capability of data, which is creating the knowledge layer, the IQ layer, the contextual layer, is the first step,” he said.

Access to transaction histories and previous service requests can help AI agents deliver more contextualised interactions than conventional systems relying on predefined rules.

Human oversight and cybersecurity

The panellists stressed that financial institutions must establish clear boundaries for AI autonomy, particularly when handling sensitive customer information or making high-impact decisions.

Gadgil said frameworks governing agent identity, authority, accountability and monitoring would be essential. While low-risk tasks could be automated, high-risk activities would continue to require human oversight.

Cybersecurity also emerged as a key concern, with AI needing to strengthen defences against emerging threats. Goel highlighted the potential of open-weight AI models to offer greater control and visibility over data usage.

He said organisations must maintain basic security safeguards, including access controls and governance, and avoid exposing sensitive information to untrusted AI models.