Inside Retail Lending: How AI and Partnerships are Shaping Credit Growth

India’s retail lending business is being reshaped by three simultaneous shifts: faster digital origination, deeper partnerships, and a growing reliance on analytics and AI across the credit lifecycle. The latest episode of The Credit Continuum by Yubi examines how lenders are adapting to that change, particularly as margins compress and loan volumes accelerate.

For a deeper dive into the conversation, the full episode is also available on Spotify.

In a conversation with Vipul Mahajan, Chief Business Officer at Yubi, Sumit Bali, Country Head - Retail Assets & Debt Management at YES Bank, discussed how traditional lending fundamentals are being reworked in a system that is now fully digital, increasingly partnership-led, and steadily moving toward AI-assisted decisioning.

Lending has changed, but the core still hasn’t

Bali traced the shift in lending from paper-led processes to real-time, data-driven decisioning. Earlier, underwriting depended on physical documents and delayed verification cycles. Today, lenders can access bureau data, banking behaviour, and income signals almost instantly.

Despite that shift, he noted that the core framework of lending remains unchanged. Successful franchises still balance three variables: margin, volume and risk. What has changed is the speed and precision with which those variables can now be managed.

Technology and analytics, in that sense, are no longer differentiators. They are basic requirements.

Partnerships and co-lending are becoming central to growth

One of the more notable shifts discussed in the episode is the rise of partnership-led lending.

For newer or challenger banks, partnerships are not optional. They are a way to access customers, data, and distribution that would otherwise take years to build. Bali described co-lending as a mutually beneficial model. Banks bring lower cost of capital, while partners bring reach and customer engagement.

That combination is already showing results. He pointed to initial outcomes where working across both secured and unsecured partnerships led to a 30% uptick in business within the first two months of going live under new co-lending structures.

The logic is straightforward. Distribution is increasingly owned by platforms and intermediaries, while balance sheet strength sits with lenders. Co-lending connects the two.

AI is moving from assistance to decision support

If partnerships are changing distribution, AI is changing how lending decisions are made and executed.

Bali outlined a shift already underway in operational workflows. Tasks that earlier took hours, such as analysing financial statements or assessing borrower profiles, can now be completed in minutes using automated systems. These systems may not take the final decision yet, but they significantly compress the time required to reach one.

The next stage, he suggested, lies in dynamic decisioning. Instead of a binary approve or reject outcome, lenders are beginning to explore whether pricing can be adjusted in real time for borderline customers, allowing more approvals while still maintaining risk discipline.

This could potentially expand approval funnels. A lender approving 35 out of 100 applications today may be able to approve more applications by aligning pricing with risk at a granular level.

The real gains are showing up in collections

While AI in underwriting gets more attention, the conversation highlighted collections as an area where analytics is already delivering observable impact.

Bali described the use of multiple scorecards across collection stages, from pre-due reminders to bucket allocation and recovery strategy. These scorecards help determine:

• which customers respond to reminders,
• which require calls,
• which need field visits,
• and when to offer restructuring or settlement.

The result is both better recovery outcomes and tighter cost control, which is becoming critical as ticket sizes shrink and margins compress.

Instead of following a fixed escalation path, collections are becoming more targeted and predictive. The emphasis is shifting from effort to efficiency.

Speed is now a product feature

Another shift is behavioural. Customer expectations around credit have changed.

Younger borrowers are more concerned with obtaining assets than they are with owning them. At the margin, speed and convenience are frequently more important than price. Bali noted that customers are becoming more willing to pay higher rates if money is accessible right away.

To adapt to these changing trends, lenders are now adopting 24x7 systems, minimising clicks, and making decisions more quickly. The key to making that experience possible at scale is automation and artificial intelligence.

Risk is becoming more dynamic and more exposed

Faster lending brings its own risks. As decision cycles compress, lenders have less time to manually intervene. That makes the design of scorecards, data pipelines, and risk frameworks more important.

Bali emphasised that the “secret sauce” lies in how data is translated into decisioning logic. Access to data is no longer the advantage. The ability to interpret and apply it is.

At the same time, digitisation increases exposure to fraud and system vulnerabilities. As lending becomes more automated, the cost of failure may also rise.

Governance and data control are moving to the centre

The conversation also flagged a broader shift toward tighter governance.

With increasing digitisation and the rise of consent-driven data frameworks, lenders may no longer be able to rely on proprietary data advantages alone. Data ownership is likely to move closer to the customer, with access becoming more permissioned and portable.

This has two implications:

• lenders may need stronger governance and compliance frameworks,
• and competitive advantage may shift toward execution, not just data access.

What it means for new entrants

For fintechs and new lenders entering the space, the advice was consistent across the discussion.

• Focus on solving real customer gaps rather than replicating existing models
• Build with technology at the core, not as an add-on
• Keep governance and risk discipline embedded from the start
• Be clear on the path to profitability, especially in a margin-compressed environment

For individuals entering the sector, the message was simpler: stay grounded in customer needs and avoid shortcuts in execution or ethics.

The direction of travel

The discussion reflects a lending ecosystem that is no longer defined by distribution alone.

Partnerships are expanding reach. AI is improving speed and decisioning. Scorecards are refining risk and collections. Governance is tightening the boundaries within which all of this operates.

The combination is pushing the industry toward a model where lending is faster, more embedded, and more data-driven, but also more dependent on how well systems are designed and controlled.

If there is a single thread running through the conversation, it is this: scale is no longer the hard part. Scaling responsibly is.

Note to the Reader: This article is part of Mint's promotional consumer connect initiative and is independently created by the brand. Mint assumes no editorial responsibility for the content.