Between the digital payments revolution, the aggressive growth of fintech lending platforms, the NBFC expansion into previously underserved markets, and the post-pandemic push to drive credit penetration deeper into Tier 2 and Tier 3 cities, the last several years produced a remarkable and largely deliberate broadening of the Indian credit landscape. Loans that would have been unavailable, or inaccessible, to a significant portion of the population a decade ago became routine. Disbursement targets were met. Growth charts moved in the right direction. Shareholder presentations looked impressive.
What the disbursement targets did not always include was a commensurate investment in what happens when the repayments stop.
And in significant numbers, across significant portfolios, they have stopped.
The consequences of that gap, between the velocity of loan creation and the rigour of recovery infrastructure, are now sitting on the balance sheets of banks, NBFCs, and fintech lenders across the country, in the form of non-performing asset portfolios that are, in many cases, larger than institutions would publicly prefer to acknowledge. The question that follows from this situation is not whether collections become a strategic priority. It already is. The question is how that priority is implemented, and whether the chosen approach protects the business, the customer relationship, and the institution’s regulatory standing simultaneously.
The answer to that question matters more than most finance teams currently appreciate.
You Should Also Read it : After 32 years in customer operations, we have been asked a version
How the NPA Problem Was Built
To understand where the collections challenge stands today, it is worth being precise about how it was constructed, because the mechanics of its creation have direct implications for the mechanics of its resolution.
The acceleration of loan disbursement in India, particularly through digital channels and NBFC networks, was in many cases characterised by documentation and verification processes that prioritised speed over rigour. Know-your-customer protocols were compressed. Credit assessment models, particularly those deployed at scale by fintech lenders operating on thin margins and high growth targets, were built on data inputs that were sometimes incomplete and sometimes optimistic. Loans were extended to customer segments whose repayment capacity had not been stress-tested against the income volatility that is a structural feature of informal and semi-formal employment patterns across much of India.
When economic pressure arrived, rising costs, employment disruption, income compression in specific sectors, the fragility of those credit decisions became visible. Accounts that had been performing moved into early delinquency. Early delinquency, without skilled intervention at precisely the right moment, moved into the later stages where recovery becomes exponentially more difficult and exponentially more expensive.
The institutions that had invested in collections infrastructure, trained teams, intelligent prioritisation, compliant processes, and the domain expertise to have recovery conversations that preserved rather than destroyed the customer relationship, managed this transition far better than those that had not. For the latter group, the NPA portfolio grew as the collections capability lagged, and the gap between what was owed and what was being recovered widened with each passing quarter.
This is where a significant portion of India’s lending sector finds itself today. And the decisions made about collections strategy in the next twelve to twenty-four months will determine not just the recovery numbers for this NPA cycle, but the institutional capacity to manage the next one.
The Myth of the Aggressive Collector
There is a persistent and commercially damaging misconception about what effective collections looks like. It tends to surface in organisations that have not spent meaningful time inside a professionally run collections operation — and it frames recovery as an inherently adversarial exercise, in which pressure, persistence, and volume of contact are the primary levers of performance.
This model is not merely unpleasant. It is counterproductive, and the evidence against it accumulates with every delinquency cycle that a sophisticated collections operation manages.
The customer who stops repaying a loan is almost never indifferent to their obligation. They are, in the vast majority of cases, in a situation — temporary cash flow pressure, a missed salary, an unexpected expense, a business disruption — that has made repayment feel impossible rather than unimportant. The collections call that treats them as a defaulter to be pursued is a call that confirms their fear that the lender does not see them as a person. It produces defensiveness, avoidance, disconnected numbers, and — in the worst cases — a deliberate escalation of non-cooperation that makes recovery harder, not easier.
The collections call that treats them as a customer in difficulty, worth understanding and worth working with, produces something categorically different: a conversation. And in collections, a conversation is the beginning of a resolution.
This is not sentimentality. It is operational pragmatism grounded in thirty-two years of experience running collections programmes across banks, NBFCs, and financial services companies. The recovery rates on accounts handled with discipline, compliance, and genuine customer orientation consistently outperform those handled with pressure and volume. The customer who agrees to a repayment plan in a respectful conversation is a customer who makes those payments. The customer who is pushed into a promise they cannot keep, extracted through persistence rather than genuine engagement, is a customer who defaults again, this time with a documented contact history that makes legal recovery more complex.
There is also a reputational dimension that institutions increasingly cannot afford to ignore. India’s regulatory environment around collections conduct is not static. The Reserve Bank of India’s guidelines on fair practices in debt collection are detailed, and the consequences of systematic non-compliance are real. In an environment where a single viral incident of heavy-handed collections conduct can produce regulatory scrutiny and reputational damage that far exceeds the value of the portfolio being recovered, the risk calculus of aggressive collection has shifted significantly.
Effective collections is not the art of extracting money from people who don’t want to pay. It is the art of having the right conversation, with the right person, at the right moment, in a way that makes paying the path of least resistance. Everything else is tactics in the service of that conversation.
Intelligence Before Intervention: The Prioritisation Problem Debt collection services
In a large delinquency portfolio, not all accounts are equal, and treating them as if they were is one of the most expensive mistakes a collections operation can make.
The account that is thirty days delinquent because of a temporary salary delay, has a strong repayment history, and has responded to every previous contact attempt is not the same recovery proposition as the account that is thirty days delinquent, has a pattern of partial payments followed by lapses, and has changed contact numbers twice in the last six months. They occupy the same delinquency bucket on the standard aging report. They require completely different approaches, different timing, different channels, different conversations, different resolution offers.
The collections operation that cannot distinguish between these two profiles at scale is an operation that will misallocate its most valuable resource, the time and attention of skilled agents, across accounts in direct proportion to their volume rather than their potential for recovery. The easy accounts that would have been resolved with minimal intervention receive the same level of intensity as the difficult ones that require genuine skill. The difficult ones that are actually resolvable, with the right approach, at the right moment, are treated identically to the ones that are not, and the conversion opportunity closes before it is identified.
Intelligent prioritisation, the ability to segment a delinquency portfolio not by aging alone but by behavioural indicators, channel responsiveness, payment propensity, and the specific circumstances most likely to determine resolution outcome, is the operational capability that separates collections performance at the top of the distribution from collections performance at the average.
This is where AI-augmented collections intelligence has delivered demonstrable, measurable value within our operations at Tele Access. Predictive models that analyse payment behaviour patterns, contact history, channel engagement, and account characteristics can identify, within a large portfolio, which accounts are most likely to resolve at each delinquency stage, and recommend the specific channel, timing, and approach most likely to produce that resolution. The human agent who makes the call does not arrive at the conversation without context. They arrive knowing what the data suggests about this customer’s situation and receptivity, which shapes the conversation before the first word is spoken.
The result is not merely better recovery rates, though the rates improve. It is a more precise allocation of skilled human attention to the moments and accounts where that attention generates the most value, and a more efficient use of AI-assisted outreach for the accounts and stages where technology can initiate contact as effectively as a human agent.
The Promise-to-Pay Problem
There is a specific operational failure in collections that is more common than most institutions track carefully, and more commercially damaging than its apparent simplicity suggests.
A promise to pay is not a recovery. It is the beginning of a recovery — and only if it is made under conditions where the customer can actually fulfil it.
The collections operation that optimises for promise-to-pay rates without tracking promise-kept rates is optimising for a metric that does not correspond to actual cash recovery. Worse, it is potentially producing a documentation trail of promises that cannot be kept, extracted through conversations that prioritised commitment over feasibility, which creates complications when the account moves toward legal recovery and the institution needs to demonstrate that it made genuine, good-faith efforts to resolve the debt consensually.
The discipline of structuring a repayment conversation around what the customer can actually do — rather than what would look best on this quarter’s promise-to-pay report — is a discipline that requires both training and institutional commitment. It requires agents who understand that their job is not to extract a commitment in this call but to facilitate a resolution that actually results in money recovered. And it requires a quality framework that tracks and rewards promise-kept rates at least as rigorously as promise-to-pay rates.
This distinction between the collections culture that optimises for conversation outcomes and the one that optimises for conversation metrics, is one of the most reliable predictors of long-term recovery performance on a large portfolio. The institutions that have built their collections function around genuine resolution rather than commitment extraction consistently recover more, at lower cost, with fewer regulatory complications, and with a meaningfully higher proportion of customers who remain viable for future lending relationships after the delinquency is resolved.
Recovery That Preserves the Relationship
The framing of collections as purely a recovery function misses something commercially significant: the customer who repays a delinquent account, treated with dignity and competence throughout the recovery process, is a customer who can be re-engaged. In lending, as in every financial services vertical, the lifetime value of a customer who resolves a difficulty and continues the relationship is substantially greater than the recoverable value of a single delinquent account.
The institution that recovers the debt and destroys the relationship in the process has optimised for a single transaction. The one that recovers the debt and preserves the relationship has recovered a customer, and in a market where customer acquisition costs continue to rise, the value of that distinction is not trivial.
This is why, across our collections practice at Tele Access, the orientation of every recovery conversation is dual: resolution of the current obligation, and preservation of the customer’s sense that the institution is one worth continuing a relationship with. These objectives are not in tension. Handled correctly, they are mutually reinforcing. A customer who feels treated fairly in their most financially vulnerable moment does not forget it.
The Case for a Specialist Collections Partner
Collections is not a function that can be staffed up quickly from a general operations pool when an NPA portfolio reaches uncomfortable proportions. The skills required — the regulatory knowledge, the conversation architecture, the delinquency psychology, the AI-augmented prioritisation capability, the quality framework that tracks real recovery rather than intermediate metrics, take years to develop and require continuous investment to maintain.
The institutions that will manage the current NPA cycle most effectively are not necessarily the ones with the largest collections teams. They are the ones with the most intelligent ones, teams that understand the difference between a customer who is unwilling and one who is temporarily unable, that can identify and act on propensity signals before accounts age beyond cost-effective recovery, and that can conduct the kind of conversation that produces genuine resolution rather than a promise that no one believes will be kept.
At Tele Access, we have been running collections and recovery operations for over ten years, across banks, NBFCs, fintech lenders, and financial services companies, with a consistent orientation toward recovery that is compliant, intelligent, and genuinely customer-first. Not because those are aspirational values. Because in three decades of customer operations, we have seen what the alternative produces. And it produces worse recovery rates, higher regulatory risk, greater reputational exposure, and a portfolio of destroyed customer relationships that represent lost future value as surely as they represent present-day write-offs.
India’s loan books are full. The repayments aren’t coming in at the rate they need to. What happens next depends entirely on the quality of the recovery conversations that follow. Those conversations deserve to be had properly
