Indian banks, insurance coverage corporations, and fintech corporations are prototyping, testing, and deploying a spread of AI-powered measures in customer-facing departments. A few of these corporations are making a brand new transition from software program consumers to software program builders.
Few individuals have been on the forefront of this transformation like Kiran Jagannath, head of economic companies and conglomerates at AWS India. In a wide-ranging dialog on the sidelines of AWS re:Invent 2025, he mentioned how cloud, generative AI, agent orchestration, information modernization, and tradition are converging into what he calls “a complete new working mannequin for monetary companies.”
From his perspective, Indian monetary companies corporations are at completely different phases of this alteration. Fintech startups are “digital natives born within the cloud” and have little to fret about legacy. Surprisingly, insurance coverage corporations had been among the many earliest to undertake cloud, beginning round 2017. “Insurance coverage in India remains to be comparatively younger,” Jagannath defined.
“They’re naturally drawn to the cloud as a result of they do not have a number of legacy debt and might leverage information to distinguish and innovate shortly.” (Legacy debt within the tech world refers to older software program and {hardware} techniques which might be unsuitable for present era AI workloads.)
one other story
However banks inform a unique story. They began their cloud journey late, however as soon as they did, they accelerated quicker than another section. Jagannath stated the turning level got here throughout the coronavirus pandemic. “When branches might not operate usually, banks wanted to broaden their digital channels inside weeks. Traditionally, tasks of that measurement took quarters and even years. The cloud has modified that.”
The second set off was the UPI explosion. Jagannath identified that the unpredictable, almost doubling annual progress has compelled regional banks to utterly rethink their procurement fashions. “You possibly can’t estimate the amount 5 years from now. You possibly can’t even estimate the amount two years from now. With this sort of volatility, autoscaling on the cloud turns into important.”
He believes the cloud downside is solved. “The query is, ‘What’s the proper structure and what’s the proper migration path?'”
This modification has paved the way in which for the following frontier: generative and agentic AI. Early use instances had been task-specific, corresponding to lowering failed funds, enhancing fraud detection, and rushing up customer support. “The long run is agentic AI,” Jagannath stated. “As a substitute of 1 agent, many brokers with human-like reasoning work collectively to finish the method end-to-end.”
He pointed to document-heavy workflows corresponding to KYC, service provider onboarding, and mortgage origination as preliminary beneficiary. “Brokers are ultimate wherever people carry out non-value-added duties, corresponding to gathering information, validating it, and sending it to a system,” he stated. “Workers will have the ability to deal with value-added actions slightly than monotonous duties.”
radical rethinking
Nonetheless, agent transformations can’t be layered on prime of older techniques. Legacy cores, monolithic apps, and inconsistent information structure are actual constraints. Jagannath acknowledged the constraints: “For an agent to be actually highly effective, the underlying techniques should be trendy.”
“The cloud is about extra than simply shifting functions from one place to a different. The true worth lies in cloud-native architectures like microservices, autoscaling, and resilient, remoted techniques. Banks are already modernizing their cores by hollowing them out or constructing twin engines.”
Tradition is simply as vital as know-how. Agent AI techniques require a elementary rethinking of how groups work. Outdated workflows are torn down and new ones constructed. Based on Jagannath, this new period will see a convergence of enterprise and know-how teams, with smaller technical teams constructing all the agent workflow.
agent world banker
When requested what a banker’s job could be like within the company world, Jagannath was cautiously optimistic. “It’s nonetheless evolving,” he stated. “Agent AI is lower than a 12 months outdated as an actual trade motion. However I do not suppose digital bankers on cell phones are far-fetched. We’re already seeing analysis assistant brokers with personalized fashions. The tempo of change is insane.”
He sees probably the most seen adjustments within the subsequent six months to be within the person expertise. Navigation-focused cellular apps are probably to present technique to voice-driven interactions, the place prospects merely converse instructions whereas an agent handles the backend logic. “Banking will turn into context-specific,” he says. “In case you’re procuring, you may test your stability or full a cost inside that have with out leaving the circulation. The banking expertise is embedded inside what prospects are already doing.”
issued – December 22, 2025 8:10am IST
