From credit score to context: How Hero FinCorp is using AI to transform lending

Priya Kashyap, Chief Operating Officer, Hero FinCorp.For years, lending decisions have been built around a familiar set of signals: bureau scores, repayment histories and financial documents.But those signals can tell only part of the story.For a lender operating across millions of customers and thousands of markets, the more difficult question is whether a borrower’s ability and willingness to repay can be understood in the context of where they live, how their income behaves, what they are buying, how they interact with the business and how conditions in their local market are changing.That is the problem Hero FinCorp set out to address.Its answer, Project Dhruv Tara, an AI-led lending architecture designed to bring intelligence across the credit lifecycle — from customer sourcing and underwriting to onboarding, portfolio monitoring, servicing and collections.

The ambition is not simply to make lending faster, but to make decisions more contextual while maintaining discipline around risk, credit costs and portfolio quality.Looking beyond the bureau scoreTraditional lending models tend to depend heavily on bureau scores, repayment history and static financial information.

While these remain important, Hero FinCorp is using a broader set of signals to build a more complete picture of the customer.Its data framework brings together bureau and repayment information with banking signals, GST data, account aggregator inputs, demographic indicators, application information, sourcing-channel behaviour, asset characteristics and geography.These signals feed predictive models that assess factors such as repayment capacity, financial stability, resilience, intent and suitability.AI allows lenders to look beyond static bureau scores and evaluate risk through a wider, more contextual set of signals — including behaviour, geography, affordability and market-level patterns.Priya Kashyap, Chief Operating Officer, Hero FinCorpThe approach is particularly relevant for customers with limited formal credit histories, where a thin credit file does not necessarily provide a complete picture of actual credit risk.Making local intelligence work at scaleA major part of Project Dhruv Tara is the ability to bring local and market-level intelligence into lending decisions.Hero FinCorp’s cloud data lake incorporates more than 100 micro-market data points and 40+ APIs, allowing the company to analyse patterns at a granular, pin-code level.This includes understanding local income patterns, asset behaviour, geography and other market characteristics that can influence repayment behaviour.At its current scale, Hero FinCorp serves more than 13 million customers across over 96 per cent of India’s PIN codes, making this level of contextual intelligence important not just for individual decisions but for managing portfolios across very different markets.From faster processing to better decisionsThe value of AI in lending ultimately comes down to the quality of decisions it enables.Hero FinCorp has deployed 15 dynamic risk and collections models across its lending operations, using predictive intelligence to strengthen risk selection and identify emerging patterns earlier.The company reports that credit costs declined 15 per cent in H2 FY26 compared with H1 FY26, while GNPA improved by around 20 per cent over the same period.Its proprietary machine learning scorecards are also being used to improve risk selection and portfolio quality.“The real value of AI in financial services is not speed alone.

It is better decisioning.

When AI helps improve risk selection, reduce friction, detect stress early and personalise customer engagement, it becomes a core business capability” says Priya.AI is also making the customer journey simplerThe use of AI extends beyond underwriting and risk.Hero FinCorp has embedded AI into customer servicing through chatbots, IVR systems, automated workflows and AI-enabled email.

These systems now handle nearly 60 per cent of customer queries, saving more than 4 lakh human minutes every quarter.The company is also redesigning customer journeys around simplicity.

One onboarding journey, for example, has been reduced from 90 fields to just three, reducing the amount of information customers need to provide.The objective is to make AI visible in the customer experience — through faster interactions, simpler processes and more relevant engagement — rather than keeping it limited to the back end.Making collections more predictiveCollections is another area where Project Dhruv Tara is changing how the business operates.Instead of treating borrowers through broad, one-size-fits-all approaches, predictive models analyse behavioural patterns, identify emerging risk and help determine when and how customers should be approached.As Priya explains, “Collections and servicing are being reimagined through AI.” The company can now understand borrower behaviour better, intervene earlier and communicate in a more timely, respectful and customer-centric manner, rather than relying on broad-based engagement.The intent is not simply to increase collection activity.

It is to make interventions more timely and differentiated, while keeping the customer relationship at the centre.

This shifts collections from a largely reactive process towards one where the organisation can identify signals earlier and respond with greater context.Keeping humans in the loopFor all the automation and predictive intelligence, Hero FinCorp's approach does not remove human judgement from lending.AI is used to process patterns and signals at scale, while teams continue to provide contextual understanding, governance and judgement — particularly in decisions where customer impact and accountability matter.That distinction is important as AI becomes more deeply embedded in financial services.

The technology can surface patterns, predict outcomes and automate parts of the workflow, but responsibility for the decision still sits with the organisation.Project Dhruv Tara reflects that broader shift — from using AI as a collection of tools to building intelligence into the way lending decisions are made.For Hero FinCorp, the next phase of AI-led lending is therefore not simply about processing more data or automating more tasks.

It is about using richer context to make better decisions, identifying risk earlier, reducing friction for customers and embedding intelligence across the entire lending lifecycle.