India’s Setu launches Sesame; a Large Language Model for BFSI sector

By Gloria Mathias

Setu, a part of the Pine Labs Group, an API infrastructure company, has unveiled Sesame, a Large Language Model (LLM) specifically designed for the BFSI sector in India.

The Indian FinTech Setu has collaborated with indigenous AI research firm Sarvam AI, which marks a “ChatGPT Moment” in the financial services space.

Sesame leverages the rich data ecosystem enabled by India’s digital infrastructure to power features such as improved credit underwriting, fraud detection, loan monitoring, upsell/cross-sell and personal finance advisory.

Setu and Sarvam AI have together created Sesame to be both domain and region-specific, by training it on custom data that is highly relevant to India’s BFSI sector. By combining this domain-specific LLM with delivery models that allow enterprise customers to harness the vast amounts of financial data available both internally and externally in a compliant manner, Setu’s vision is to empower BFSI customers to make smarter, faster credit decisions and provide hyper-personalised financial services to their customers across their entire lifecycle.

“At Setu, our mission is to democratise financial services,” said Nikhil Kumar, Co-founder of Setu. “We believe that the powerful combination of the Account Aggregator framework and transformative technologies like Large Language Models (LLMs) is the key to making this a reality. Just as India has led the way with groundbreaking innovations like Aadhaar and UPI, we are excited to build meaningful, compliant, and responsible AI applications that showcase India’s leadership in Applied AI,” added Nikhil Kumar.

“Generative AI represents a significant step change in the nature of computation. What can be done with one rupee of compute is now dramatically more valuable,” said Pratyush Kumar, Co-founder of Sarvam AI.

“Our collaboration with Setu is an example of bringing this technology to create value in the BFSI space. We showed that we can significantly increase the accuracy of identifying insights from financial statements. This is a template which we believe can be repeated across sectors where India has created foundational digital public goods,” Pratyush added.