BFSI sector in no hurry to adopt complete automation: Perfios Group CEO

The banking, financial services and insurance (BFSI) sector is in no hurry to adopt complete automation or artificial general intelligence (AGI), suggests Nitin Chugh, managing director and group CEO, Perfios, owing to the conservative and highly regulated nature of the sector.

“Financial services, by nature, are very conservative. Also because it's regulated as it should be. And therefore, I don't think financial services would be in a hurry to say that AGI is the next big thing on the anvil. Let's adopt it. Let's start using it... That's not likely to happen,” he told the Economic Times Digital during the Global Fintech Fest (GFF) 2026.

Chugh also believes that a human will never be out of the loop even if we achieve large levels of automation except for some low-level, basic tasks that are repetitive. However, automation will help an underwriter take decisions on more loans in a day.


Also Read: Your bank account could one day become an AI agent: Perfios Group CEO Nitin Chugh

He added, “Maybe for some very, very low level tasks, which today also are not very clear as to which ones don't need a human at all, but I don't think we foresee that sort of a situation where you don't need humans at all.”

The business-to-business (B2B) SaaS fintech launched its agentic AI operating system at the GFF, which is designed to help banks evaluate credit based on everyday real-world data, especially for rural India, and micro, small and medium enterprises (MSMEs).

The autonomous system, as the firm claims, turns daily dairy collection payouts and UPI records into a trusted credit profile, giving rural farmers direct access to formal bank loans; combines GST, trade, and banking data of MSMEs to help them search and apply for government funding schemes like PM Mudra (PMMY) and PM Vishwakarma.

It also analyses earning, saving, and spending habits of young customers to offer them personalised budgeting advice and timely investment guidance.

Explaining how the operating system stitches together a bunch of agents, which perform activities such as signal collecting, Chugh stressed on the importance of AI in dealing with unstructured data. “Processing unstructured data with human intelligence requires a different level of learning... because how do you create knowledge bases based on things which are not similar, even one version is not similar to the other. So, that's where AI does a better job.”

AI & the knowledge graph

Chugh said Perfios’ operating system helps widen the knowledge graph for an underwriter allowing them to make more informed decisions. “With AI... you can look at something like, let's say, what is the overall climatic pattern of that village? Is it prone to flooding or droughts? Or does it always have a delayed monsoon? Or is the nearest mandi very far away from there?... Do they have a primary school? Do they have a primary health center? Now, that gives you a good indicator of the overall wellness of the village.”

Also Read: GFF 2026: Axis Bank, J.P. Morgan see banking move from AI that knows customers to AI that acts

The system, as per Chugh, then brings together the responses to create a comprehensive risk profile of the customer.

He added: “You can use satellite imagery, you can do multiple other things, bring all of that together, there are public databases that you can go to... create a much better knowledge graph of that village, which as an underwriter... I can make a very informed decision to say, yes, there is a borrower in that village, and that village has a higher risk or a lower risk.”

Automate only where necessary

Krishna Chaitanya B, chief product officer, Perfios, believes one must be cognizant of their AI usage as it requires a lot of computing power. “We are basically not using a sledgehammer to kill an ant. You can't use the power of LLM to say that, 'Hey, you know what? Automate this process.' You're burning GPUs... every time you are training your own SLM, it costs a lot of money. It doesn't come for free.”

Chaitanya also stressed that Perfios employees use an AI agent only when it can add value over the learned data and models they have. He adds that running agents alongside existing teams creates “two separate maintenance tasks... you actually increase the labour, increase the cost as well as increase the manpower”.

But, what happens when one automates the critical tasks in an underwriting or risk assessment process? Does the human presence remain?

He suggested the presence of a human is a function of the risk appetite of the lender and not technology, and that larger financial institutions such as banks would still want the human in the loop.

Also Read: BFSI moves beyond AI pilots as firms chase measurable business value: Report

“There is a small NBFC and an aggressive NBFC... saying, this is a market which is untapped, I'm going to put [money] on it. Think about blue-collar workers... And then... [they may not need a human]. If it's a big bank, on the other hand, that big bank will say, you've done a great due diligence, I know I'm burning money, this is a new segment I'm opening up, but let me burn the money. I'll send a human being down there to see if the shop actually exists or not... I know the video evidence is there but I would still vet it.”

So, which jobs will eventually vanish with AI?

Chaitanya said that AI will absorb jobs that run around a single variable he calls ambiguity. According to him, ambiguity refers to the room a task leaves for judgment once the data runs out. Where a decision can be made from complete information and a fixed workflow, he argues, an agent will take over fastest; where judgment is unavoidable, human roles recede further into the future, a five-year horizon rather than a two-year one.

“Agents come into the picture where there is ambiguity, and where there is the lowest amount of ambiguity, the replacement will be the fastest,” he said.