Tech Explained: AI makes banking tech affordable for smaller institutions: Infosys Finacle CEO in Simple Terms

Sajit Vijayakumar, Chief Executive Officer, Infosys Finacle.

As banks accelerate digital transformation amid regulatory complexity and macro uncertainty, Infosys Finacle, a digital banking solutions player, is embedding AI across its product stack.

In this conversation, Sajit Vijayakumar, CEO, Infosys Finacle, explains how AI is reshaping Finacle’s offerings, where the company sees opportunity, and what the shift means for discretionary technology spending.

During the Q3FY26 earnings conference call, Infosys CEO mentioned that Finacle will be more AI-oriented. Can you elaborate?

We tell our customers that Finacle will help engage, transform, innovate better, and operate better — broadly, the boundaries of what a bank would do. The digital experience studio is an example of innovating and engaging better. You can roll out many customer experiences quickly and accurately.

We have developed a small language model (SLM) that understands Finacle’s potential defects from the past, for which there are solutions. This is like a fully trained support analyst, and more knowledgeable than any person. So, when an end-of-day job runs longer than expected, the customer can type in the issue and find out the solution. This is called a SupportPro and usually has about 700 live users. One no longer has to reach out to the concerned person and can instead quickly find a solution.

When transforming or upgrading Finacle from one version to another, customisations may be done by the customers, us, or anybody else. Now, an AI agent reads those customisations, understands and codifies them in a machine-understandable form, and regenerates the code in the new version.

For instance, with an AI assistant, you can initiate a $10,000 deposit with a specified maturity, and the system will auto-fill any missing data for review and submission — eliminating the need for Finacle training and potentially delivering over 50% efficiency gains. We conducted a formal study with some clients in Australia, and they reported over 50% gains in employee productivity driven purely by AI.

You can also generate code by describing requirements in plain English — the agent interprets them and writes the code, which is then reviewed, tested, and validated before deployment. We are seeing AI drive positive impact across customer interactions as well as internal product build.

With AI and Cowork plugins raising concerns about disruption in software, do you see Finacle as insulated from this shift?

We are on the other side of the table. We have a fully AI-powered experience studio with 35 or so agents. Finacle has been traditionally providing software for various types of composable banking. However, each client would try to build that user journey themselves because it is core to their business. We, being a software provider, would just support them as part of a long journey. Essentially, we deploy the software, and they’ll design their journeys themselves.

We also have an agentic experience studio where one could make a user design externally, drop it in, and it will autonomously get converted into working code. This becomes a mobile banking app. You can build the app with all the banking capabilities with around 80 per cent efficiency. It allows banks to launch multiple-user experience journeys for different customer classes, all made possible because of GenAI advancements. We benefit substantially because the utility of the product has increased multiple times now that it is AI-embedded. We are rejoicing that AI is doing what it is doing as opposed to seeing it as a pain.

As AI innovation accelerates and client expectations rise, do you see banks actively integrating these advances into their processes, or are they proceeding cautiously given regulatory constraints?

AI adoption will happen well. It is about how you design this upfront. For example, if I have to predict which ATMs are likely to run out of cash, there is no consequence if I get it wrong. However, if the AI generates a wrong tax certificate and sends it to an end customer, it creates a lot of heartburn. Depending on the function, Finacle has different tools.

We have a predictive AI tool that will tell you which ATMs are likely to run out of cash. A GenAI tool that understands human conversations well. However, to generate a tax certificate, you will force that AI to call a program written specifically for it. So, you put a deterministic logic. To go somewhere in between, you use an SLM and not an LLM because the latter does not hallucinate beyond a point and sticks within your boundary. So, between a predictive AI, an LLM, SLM, and a deterministic logic, you decide what is fit for purpose. By mixing and matching this, regardless of the regulations, you will see more AI adoption.

Based on your client interactions amid recent uncertainty, do you see any signs of a revival in discretionary spending?

Currently, customers are spending cautiously due to market dynamics. That being said, the whole of AI and transformation is indeed discretionary spending because then people deploy more products and expand to more markets. AI has reduced rollout costs.

While an existing customer may now spend slightly less per deployment, the lower cost base is expanding the addressable market. Banks that previously could not afford such platforms, especially smaller institutions, can now adopt them. What was economically unviable five years ago has become both accessible to smaller banks and profitable for us.

We are seeing AI drive positive impact across customer interactions as well as internal product build.Sajit Vijayakumar, CEO, Infosys Finacle

Published on February 23, 2026