“Everyone can pilot AI; Few can scale it”: Godrej Capital CTO Jyothirlatha B on making AI pay off

Indian financial services has no shortage of AI pilots. Far fewer lenders can show what those pilots did to cost, speed and conversion. Godrej Capital is trying to be one of them.

Jyothirlatha B., the company’s CTO, is blunt about the difference. “Our philosophy is simple: everyone can pilot AI; few can scale it,” she says. “We therefore evaluate AI through its business impact rather than the number of experiments we run.”

A lender where technology is part of the model

Godrej Capital is the financial services arm of the Godrej Industries Group and the holding company for Godrej Housing Finance and Godrej Finance. It is a relatively young lender. Its leadership has set ambitious growth targets, and has said it plans to build digital lending infrastructure with Salesforce to speed up loan processing. Growing at that pace without cost and control problems is partly a technology question.

Jyothirlatha’s answer starts with ownership. “What differentiates us is that nearly 90% of our technology capabilities are built in-house, giving us greater agility, stronger ownership of our platforms and the ability to rapidly respond to evolving business and customer needs,” she says. “This has enabled us to scale efficiently while maintaining strong governance and control.”

She describes technology as “a key business enabler across our lending and investment businesses,” with “meaningful improvements in turnaround times, operational efficiency, employee productivity and customer experience through digitisation and automation.”

Process first, then automation, then AI

Many CIOs will recognise her sequencing. She argues that AI should come last in the chain, not first.

“Across financial services, we are moving from digital-first to increasingly intelligent, data-led journeys,” she says. “In lending, technology is integrated across the lifecycle, from lead generation and onboarding to decisioning, servicing, collections, and post-loan engagement.”

The gains, she says, go beyond automation: “better use of data is enabling sharper customer insights, personalisation, stronger risk assessment, and more efficient customer acquisition.”

Then comes the line that reads like a rule for the team: “Our approach is to first optimise the process, then automate it, and layer AI where it creates demonstrable business value. Technology should solve a problem, not become the solution looking for one.”

Other Godrej group companies make a similar argument. Godrej & Boyce has said it stresses digitising processes before applying AI, so that data collection is consistent and reliable.

AI inside the engineering shop

The more unusual claims concern how Godrej Capital builds software. “In engineering, 30% of new code is AI-written, contributing to a 45% improvement in delivery velocity and 50% lower cost of delivery,” Jyothirlatha says. “We are also using AI across back-office operations, customer service, regulatory analysis, and reporting.”

She also pushes back on the idea that this is only about tooling. “Technology transformation is equally about people. We consciously create space for our teams to experiment, build, and own solutions, giving employees a strong sense of ownership and entrepreneurial spirit. We want our people to solve real business problems and create impact, rather than simply execute technology projects.”

Saksham and the AI Lab

Much of the AI work runs through Saksham. Earlier descriptions call it an enterprise-grade GenAI and ML platform that centralises AI development, enables secure LLM integration and simplifies governance. Jyothirlatha now describes it as “our Enterprise AI Foundry developed by the AI Lab, which provides AI tools, agents, and services through an AI marketplace.”

Her headline production example is onboarding. “One example is our AI-developed Personal Loan onboarding platform, which has delivered a 90% reduction in cost,” she says.

Her other examples are in customer targeting and cross-sell: “In our mass-market businesses, data-backed capabilities are helping improve the timing and relevance of customer outreach, while in cross-sell they are helping strengthen lead quality, predict customer intent, and improve conversion efficiency.”

She is careful about how she measures all this. “Importantly, we evaluate AI through business outcomes, not standalone technology metrics: conversion, approval rates without compromising portfolio quality, acquisition costs, efficiency, and portfolio performance.”

Build where it differentiates

Building in-house does not mean building everything. In July 2025 Godrej Capital announced a collaboration with Salesforce, with Deloitte implementing, to bring an AI-driven loan origination system to its subsidiaries. That fits a common pattern: buy the commodity layers, and build the ones that differentiate.

Next: AI-first, then agentic

Jyothirlatha frames the next phase as a change in operating model. “The larger shift we see is from digital-first to AI-first financial services,” she says. “AI will increasingly become embedded across customer journeys, decision-making, and operations. In lending, that means acquisition, onboarding, credit, servicing, and other operational workflows.”

On the investments side she sees opportunity in “RM productivity, customer intelligence, proactive portfolio insights, and more personalised engagement.”

The most ambitious claim is about agents. “The next frontier will also be agentic AI, with increasingly autonomous workflows,” she says. “Our FY28 roadmap, for instance, includes moving towards 80% of back-office processes being managed by AI agents, greater AI-led customer service, and more autonomous workflows across enterprise functions.”

She frames this carefully. “These are future-state ambitions rather than current achievements.”

Governance as the constraint

Regulation shapes how far autonomy can go. In August 2025 the RBI released its FREE-AI framework, which sets out guiding principles for ethical AI adoption in finance. Institutions would need board-approved AI policies, AI-specific risk reviews before product launches and customer disclosures when they interact with AI systems. The framework is advisory for now, and adoption is likely to be uneven until it is codified in binding directions.

Jyothirlatha’s position fits well inside that. “Greater autonomy must come with stronger accountability. Human judgement, governance, security, and AI literacy will remain fundamental. Our view is that AI should augment expertise rather than replace it.”

She ends with a view of who will win: “Ultimately, the institutions that create the greatest value will be those that successfully combine AI and data intelligence with human judgement and empathy.”