Fintech lending, BNPL, and unsecured consumer loans are driving the rapid growth of India's credit system. Despite advances in underwriting and digital disbursal, collections lag behind, depending on fragmented agency networks and operations requiring much manpower.
During his conversation with The Economic Times Digital, Ananth Shroff, co-founder and CEO of DPDzero, shared insights on establishing an AI-focused collections infrastructure, the critical role of empathy and compliance in recovery processes, and how the expansion of unsecured lending in India is influencing the future of collections. Edited excerpts.
Economic Times (ET): What led you to start DPDzero, and what gap did you see in India’s debt collections ecosystem?
Ananth Shroff (AS) : I was at the fintech startup Setu, on the cusp of leading the goal of one billion transactions through BBPS, when the acquisition happened. By then I had already been mapping the credit value chain closely - and a pattern had become impossible to ignore.
Capital in the ecosystem was abundant. Digital acquisition was exploding. Underwriting had become a closely guarded science that every lender treated as their secret sauce. And then there was collections, the ignored stepchild of lending.
No one cared. No one bothered. The lending stack had modernised dramatically, growing at a 132% CAGR over the preceding years. Collections had not moved an inch. It was still running on a traditional agency network built on zero borrower intelligence, high-intensity aggression, and standards that varied wildly from one agent to the next.
What troubled me was not just the inefficiency. It was the absence of any structured, technology-led platform that could scale recovery in a compliant, consistent manner. A large part of the ecosystem still relied on volume calling, unannounced field visits, and approaches that bore little resemblance to how modern credit behaviour actually works. Collections have been running on a dignity deficit without anyone taking notice.
There was also a deeper perception problem. Collections have historically been treated as the uncomfortable last mile of lending-difficult, unorganised, and best outsourced and forgotten. But that framing gets it completely backwards. To me, collections is fundamentally what enables the entire lending ecosystem to function sustainably with confidence.
DPDzero was built to fix that, to change the way how India collects. Our goal is to ensure that every loan disbursed in this country has an intelligent, empathy-first recovery path behind it. That gap was real, it was large, and frankly, it was waiting for someone to take it seriously.
ET: When was DPDzero founded, and how has it evolved since then?
AS : DPDzero was founded in 2022 with the idea of reimagining collections from a fragmented outsourcing function into a unified, intelligence-driven infrastructure layer for lenders.
In its early days, the focus was on building structured digital communication workflows and improving consistency in borrower engagement. However, as lending volumes grew and borrower complexity increased, the platform quickly evolved into a full-stack Collections-as-a-Service (CaaS) model.
In fact, within one-and-a-half years of operating as a SaaS, we realised that what lenders need is a trustworthy partner who can run end to end collections for them while giving them a complete view of what’s happening on their portfolio. That is when we took a decision to pivot to CaaS.
Today, DPDzero operates as an AI-led platform that integrates decision-making intelligence, borrower segmentation, multi-channel orchestration layer, and a robust in-house TeleCalling and Field network empowered with various AI capabilities.
The key differentiator is that our frontline agents (or any engagement channel with the borrower) have complete borrower context all the time, making communication effective and empathetic. At no point - be it a telecaller in front of a screen or a field agent with our field app at the doorstep of the borrower-the borrower context is never lost.
Over time, the platform has expanded significantly in scale. DPDzero now manages over Rs 1,400 crore of portfolio per month, impacting close to 20 lakh borrowers monthly, and has processed more than Rs. 6,000 crores in collections to date. Growth has been driven by increasing demand from lenders seeking structured, compliant, and technology-led alternatives to traditional agency models.
ET: How has the unsecured lending boom shaped demand for AI-led collections?
AS: The unsecured lending boom in India has not just expanded credit access, it has fundamentally rewritten the collections problem. The growth which is driven by fintech credit, BNPL, and consumer lending has seen a 13.5% increase in volume in the recent fiscal year.
Unsecured personal loans under Rs 50,000 have surged by 48%. The volume is staggering, and the average ticket size is small, which means the cost-to-collect math is brutal for anyone still relying on manpower-heavy models.
But volume is only half the problem. The borrower profile has also shifted dramatically. Over half of the fintech unsecured loans today go to borrowers under 35 - first-time credit users, gig economy workers, borrowers who have never interacted with a recovery agent and expect the same digital experience in collections that they got at disbursal.
Traditional call-centre-led recovery models were not built for this cohort. They break on consistency, compliance, and cost simultaneously. They are increasingly difficult to scale efficiently.
This is precisely where AI-led collections stop being a nice-to-have and become structurally necessary. AI enables lenders to run intelligent, personalised outreach at scale - across millions of borrowers, across channels, and DPD buckets - without the unpredictability of human-led operations. It puts consistent performance on repeat mode.
At DPDzero, we have seen this firsthand with lenders who deploy AI-driven communication early in the delinquency cycle and see significantly better outcomes - both in recovery rates and borrower experience.
The unsecured lending boom did not just create demand for AI-led collections. It made it inevitable. And we are happy to lead that change across the credit landscape in India.
ET: What insights have emerged from 10 million+ borrower interactions?
AS: We interact with approximately 2 million borrowers every month across delinquency buckets. At that scale, the learning curve stops being linear - it becomes complex, exponential, and genuinely humbling. With those many interactions happening m-o-m, the insights have not stopped surprising us.
The single most consistent finding is this: borrower behaviour is deeply contextual, not sequential. Traditional collections logic assumes that more touchpoints mean more recoveries. Our data says otherwise. In a significant number of cases - timing, tone, and channel selection outperform outreach volume by a wide margin.
Borrowers who go silent on repeated reachout often engage meaningfully through a well-timed WhatsApp message or a structured voice interaction scheduled at the right moment in the week.
The second insight is one that the industry has been slow to absorb: most defaults are not always purely financial - they are often behavioural or situational. Temporary liquidity gaps, communication fatigue, or lack of clarity around repayment options frequently contribute to delayed payments.
A borrower who does not understand their options to pay is not the same as a borrower who refuses to pay. Each borrower is unique. Treating them identically is both inefficient and unfair.
What this demands is a platform that reads borrower context continuously - not one that treats every interaction as a fresh, isolated event. AI-driven segmentation and decisioning make this possible at scale.
The goal is not just higher recovery rates. It is recovery that does not cost the borrower their dignity or the lender their relationship with the borrower.
ET: How does your AI Decision Intelligence layer work?
AS: Most collections platforms tell you what happened. DPDzero's AI based decision Intelligence layer tells you what to do next - and why.
At the core, its primary function is to determine the most effective way to engage with each borrower - who should be contacted, when they should be contacted, and through which channel.
The system evaluates multiple parameters, including credit risk signals, repayment history, behavioural response patterns, and prior interaction outcomes. Based on this, it dynamically assigns communication strategies across SMS, WhatsApp, IVR, Voice AI, human agents, and field teams alone or combined. It gathers insights in real-time to optimise outreach strategy for maximum recovery and zero borrower inconvenience.
No two borrower journeys are identical, and the platform is designed to reflect that. A first-time defaulter with a strong repayment history gets a very different engagement sequence than a repeat delinquent borrower with demonstrated avoidance behaviour.
A key strength of the platform is its ability to learn continuously. Every interaction - a message opened, a call declined, a payment made three days after a WhatsApp nudge - feeds back into the model. As borrower responses evolve, the system adapts its decisioning in real time, improving precision in both engagement and recovery strategies.
Importantly, this intelligence layer also plays a critical role in ensuring borrower dignity and controlled communication intensity, preventing over-aggressive outreach and enabling more contextual, respectful engagement aligned with the regulatory expectations.
ET: How do you balance compliance, empathy, and efficiency in collections?
AS : I have heard a lot about collection agencies and companies struggling to find a balance between these three but for us they are an integral part to building a collections infrastructure for the evolving Indian credit economy.
Given the regulatory frameworks, RBI's Fair Practices Code, its Responsible Business Conduct Directions, and the evolving regulatory landscape compliance is a non-negotiable.
At DPDzero, compliance is not a checklist that sits at the end of the process. It is embedded into platform architecture from the ground up; governing contact frequency, communication timing, and escalation logic. Whether the interaction is AI-driven or agent-led, the system enforces compliance uniformly. There is no escaping that guardrail.
Talking about empathy, it is the moat that drives our vision. We have always believed that there is a human behind every collection, then why are collections inhuman?
That philosophy is at the core of our existence and leading the change. This pushes us to bring together the power of AI that makes the system more intelligent, context-aware and keeps everyone informed about the borrower all the time making conversations empathetic and conversions seamless.
We believe collections must evolve from being a transactional recovery exercise to a structured engagement process that respects borrower dignity. AI plays a critical role here by ensuring that communication is context-aware, non-repetitive, and appropriately paced.
And when empathy and compliance are baked into the entire stack, there is no reason for us to be concerned about efficiency, it follows on its own. We are looking to deliver a collections platform that is ethical, efficient at scale & where recovery efficiency and borrower experience are not in conflict, but designed to coexist.
ET: How has the business progressed financially, and what is your model?
AS : Our model is simple and deliberately so: we earn only when lenders recover. No recovery, no revenue. That outcome-linked structure - Collections-as-a-Service - keeps us honest, keeps us aligned with lender portfolios, and keeps us hungry to build better.
On growth, DPDzero has scaled 2X in the last one year. We work across NBFCs, fintech lenders, and have recently onboarded some of India's top private banks. The demand signal is consistent: lenders at every tier are looking to replace fragmented, manpower-heavy recovery operations with intelligent, scalable infrastructure.
We are in a deliberate high-investment phase right now - building deeper AI systems, expanding PAN India operational capabilities, and strengthening the technology foundation. We are on our trajectory to putting collections on autopilot for lenders across India.
The long-term direction is clear: building a scalable, AI-led collections infrastructure layer for India’s credit ecosystem and to collect every loan that is disbursed.
ET: Where do you see the biggest opportunity for DPDzero going forward?
AS: The biggest opportunity ahead of us is market consolidation and it is already underway.
Two forces are accelerating it. First, regulation. The RBI's expectations around compliant, structured, borrower-respectful collections are only getting sharper. The incumbent agency infrastructure was never built for this environment. It cannot absorb these requirements without a fundamental rebuild. DPDzero was built compliance-first from day one - that is not a feature we added, it is how the platform was designed. That gap between what regulators expect and what legacy players can deliver is where our opportunity lives.
Second, borrowers are changing. The next generation of credit users is digitally native, more aware of their rights, and one public post away from putting a lender's brand reputation at serious risk.
Lenders are beginning to understand this, and they are looking for partners who can deliver that change in the most seamless way possible.
My honest view: 90% of incumbent collections agencies will not exist five years from now. The industry will consolidate around four or five technology-led players who can deliver recovery at scale without sacrificing compliance or borrower experience. We are already seeing large lenders move in that direction.
The future of collections in India will not be defined by who calls the most or pushes the hardest. It will be defined by who understands the borrower best and acts on that intelligence with precision. That is the company we are building at DPDzero and we have a meaningful head start.
