Arundhati Bhattacharya on Why India's AI Moment Is Different This Time

Few Indian tech executives carry the kind of cross-industry credibility Arundhati Bhattacharya does. Before she became president and CEO of Salesforce South Asia in 2020, she was the first woman to chair the State Bank of India, the country's largest bank. That banking background, she has said in interviews, gives her a user's-eye view of enterprise software rather than a purely technical one — and it shapes how she talks about where India's AI adoption is actually headed.

From Experimentation to Execution

In recent interviews, Bhattacharya has been direct about the shift she sees happening in Indian boardrooms. Where 2025 was largely a year of businesses experimenting with individual AI agents, she has described the years ahead as a move toward coordinated, multi-agent deployments — systems where AI agents stop being narrow task-takers and start owning outcomes across functions. Salesforce has been building its India strategy around that shift, recently adding Hindi-language support to its voice-based AI agent product and preparing to let Indian enterprises host AI workflows locally through MuleSoft on Hyperforce, a response to growing regulatory focus on where enterprise data physically lives.

Candid About the Adoption Gap

Not every conversation has been triumphant. In one interview, Bhattacharya acknowledged that Indian companies have been more cautious about large-scale AI adoption compared to businesses in the US, and that most of Salesforce's India customers piloting agentic AI are homegrown conglomerates rather than multinational subsidiaries, which tend to run their AI pilots out of headquarters elsewhere rather than in India. She has pointed to sectors like banking, financial services, insurance and contact centres as the clearest early opportunities for agentic AI to improve customer experience.

Trust as the Real Bottleneck

Across multiple interviews, one theme repeats: trust, not capability, is what is actually slowing enterprise AI adoption in India. Bhattacharya has argued that AI systems cannot function as a black box — they need to be explainable and designed to support human judgment rather than replace it, with strong data governance and human oversight built in from the start. She has framed that discipline as a deliberate trade-off for Salesforce in India: being willing to move slower to market in order to guarantee that customer data adoption questions are resolved first.

A Banker's Lens on a Tech Career

Bhattacharya has also been open about how differently she now works. She has described using Salesforce's own AI-powered Slack assistant to summarise incoming messages, flag what needs urgent attention, and even help decide which meetings on her calendar are worth attending — a small, personal example of the same agentic AI logic she is trying to get Indian enterprises to adopt at scale. She has also joked about needing a "translation" of how younger engineers talk about technology, a reminder that even a company's own CEO is not immune to the pace of change she is asking her customers to keep up with.

Why It Matters

Bhattacharya's interviews offer a useful read on enterprise AI in India precisely because she isn't just selling the technology — she spent decades on the other side of the table as a banking regulator-adjacent executive who had to trust vendors with sensitive data. Her repeated emphasis on explainability and governance, rather than raw capability, suggests that the next phase of India's AI adoption curve will be won less on features and more on which vendors can prove their systems are safe to hand real decisions to.

CEO Interviews

At www.elevatexstudios.com / ElevateX Studios, CEO Interviews, part of our Interviews series, brings you conversations with the executives leading India's biggest technology, finance and consumer businesses — what they're building, what's working, and what still worries them.