Mahesh Ramamoorthy, the Chief Information Officer (CIO) of YES Bank, has shared a new vision for how the bank will use technology. He recently discussed the potential of using homegrown Indian AI models to improve operations. This comes at a time when 'tokenization costs' (the fees paid to use AI services) are a major concern for banks. By using local models, the bank hopes to solve specific problems more cheaply and efficiently.

One of the biggest hurdles in AI today is the cost of tokens. In simple terms, tokens are the small pieces of data that AI models process, and companies are charged for every token used. Ramamoorthy explained that while global 'frontier' models like ChatGPT are powerful, they are very expensive. For tasks that are specific to India, such as document extraction or converting local speech to text, the bank does not need a costly global model. Instead, local options like 'Sarvam' or 'BharatGen' can do the job better because they are 'Indianized' and cost ten times less.

YES Bank is open to using these Indian models as long as they provide clear value. However, the CIO emphasized that 'guardrails' (safety rules and security measures) are the most important factor. The bank must ensure that data privacy is never compromised when using new technology. While the maturity of these local AI models is still growing, Ramamoorthy believes they will eventually lead to much better efficiency for the banking sector.

Many people wonder if AI can quickly fix 'tech debt' (the cost of maintaining old, outdated computer systems). Ramamoorthy warned that this will not happen overnight. In the banking world, you cannot change systems with a 'snap of a finger' because banks are highly regulated. He expects that using AI to truly modernize the bank's ecosystem will take another 12 to 18 months. Safety and careful testing come before speed in banking technology.

The role of the CIO is also changing. AI is now a central part of boardroom strategy discussions. The board of directors now expects the management to weave AI into the bank's long-term plans. The goal is to show clear value generation and transform how the bank operates. It is no longer just a technical tool; it is a business strategy tool.

Currently, AI in the BFSI (Banking, Financial Services, and Insurance) sector is mostly used behind the scenes for risk management and operational efficiency. However, the next step is moving AI to the 'frontline' to interact directly with customers. This might start as 'assisted AI,' where a human helps the process, before becoming fully automated. Trust remains the most important factor, especially as the RBI (Reserve Bank of India) has released new expectations for model governance.

To help this transition, YES Bank recently held an 'AI Fest.' This event was designed to give product managers and business teams a look at what AI can do. The goal is to build a 'pipeline' of ideas where different departments collaborate to find new ways to use AI. By doing this, the bank ensures that technology is not just forced on staff, but is used to solve real business problems identified by the teams themselves.

For bank officers and aspirants, this shift means that understanding AI will become a core skill. As YES Bank makes AI central to its strategy, the focus will be on meaningful progress in areas that improve customer experience and lower costs. The banking industry is watching closely to see how these localized Indian models perform compared to global giants.