AI has reduced the number of workshops needed to move prospective banking clients through Infosys Finacle’s sales process, while the company is using smaller and locally deployed models to control costs and protect sensitive data.
Artificial intelligence is already shortening one of the most time-consuming parts of banking technology sales, according to Sajit Vijayakumar, chief executive of Infosys Finacle, who said AI has reduced the effort needed to move potential clients through long pursuit cycles.
Speaking to Business Standard, Vijayakumar said work that once took around six months and as many as eight workshops can now often be completed in two or three sessions, cutting the sales journey by 60% to 70%. He said that has significantly lowered friction in the sales process as banks become more comfortable with AI-driven tools and demonstrations.
Infosys Finacle, the banking software arm of Infosys, has been widening its use of AI across its platform, with applications ranging from operational efficiency to more dynamic pricing and personalised customer services. The company has also been positioning its Data and AI Suite as a way for banks to build predictive and generative AI tools with greater transparency, explainability and governance, reflecting the sector’s cautious approach to deploying automation in sensitive environments.
Cost remains a major hurdle. Vijayakumar warned that poorly designed AI workflows can quickly become expensive because of token consumption, the unit that measures the amount of text processed by language models. He said a single automated transaction could cost as much as $3 to $5 if banks are not careful about how AI is used. That has pushed Infosys Finacle towards a layered model strategy, in which smaller language models handle simpler tasks while larger models are reserved for work that needs broader context or deeper reasoning.
The company is also using locally deployed open-source models so that sensitive banking data does not need to be sent to the cloud. Vijayakumar said the software can choose the most suitable model in the background, depending on the complexity of the request, in order to balance cost and performance. Infosys Finacle says it has built a repository of around 200 AI skills that can operate as agents within its banking framework, although Vijayakumar said he would prefer fewer agents if the functions could be made more precise.
Beyond internal efficiency, the company sees a commercial case for AI in payments and customer engagement. Vijayakumar said payment activity can reveal patterns that help banks understand what customers are likely to buy next and enable more targeted offers. He said some US firms track as many as 800 characteristics of a person to keep refining those insights. That direction fits with broader product work at Infosys Finacle, which recently highlighted a modernisation project with HDFC Bank and a wider push towards automation-first banking systems designed to streamline operations and support more personalised services.
Disclaimer: This article is intended to inform and educate, not to recommend or endorse any financial product, investment or strategy. Please consider your own financial circumstances and seek professional advice where appropriate before making financial decisions.
