thumb
  • FrenzoCollect

  • 21-07-26

Why lenders are moving on from the debt collection company model

For thirty years, the standard answer to rising delinquency in Indian lending was to call a debt collection company. The model was simple and the logic felt sound: collections are unpleasant, agencies specialise in unpleasant, pay them on results and move on.


That era is ending, and not because agencies got worse. Three forces converged on the model at the same time.


The regulator closed the accountability gap

The old, unspoken appeal of outsourcing was distance. If recovery got ugly, it was the agency's people doing the ugly part. The RBI has spent the last few years dismantling that distance piece by piece. The August 2022 circular made regulated entities responsible for their recovery agents' conduct. The Digital Lending Directions extended the same logic to fintech-originated loans. And the draft directions published in February 2026 consolidate it all: due diligence on agencies, training obligations, grievance redressal, restrictions on what borrower data can even be shared with agents, and full accountability sitting with the lender.


The result is that outsourcing now carries most of the compliance burden of in-house collections plus a monitoring problem: you must supervise conduct you cannot directly see. Several large NBFCs manage over a hundred third-party agencies. Auditing that many field operations to RBI standard is its own department.


The economics stopped working at scale

Digital lending changed portfolio shape. Where an NBFC once had thousands of large secured loans, it now has lakhs of small-ticket unsecured ones. Agency economics were built for the first shape. A commission-per-recovery model on a ₹15,000 personal loan barely covers the agency's dialer time, so agencies respond rationally: they cherry-pick the accounts most likely to pay anyway and let the rest age.


Meanwhile the pool being handed over keeps growing. India's PAR runs around 7% of a ₹198 lakh crore credit market - about ₹11 lakh crore at risk, above the world average of 5%. Lenders throwing that volume at commission-based agencies are paying an escalating tax on their own delinquency, with the fee highest exactly where recovery is hardest.


Technology removed the reason to outsource

The honest historical case for agencies was capability: they had the dialers, the trained callers, the field networks, the process discipline that lenders lacked. That capability gap is what collections technology closed. AI scoring now identifies stress before an EMI misses, in the window where recovery probability is above 80%. Routing engines send each account to the cheapest channel likely to resolve it. Compliance guardrails run in software instead of depending on an agent's mood at 7pm. None of this requires a third party; it requires a platform.


This matters most for prevention, which is the part of collections no agency will ever sell you. An agency's revenue starts when your account goes bad. A platform's value starts before it does. On FrenzoFinserv, the biggest single recovery uplift lenders see is in the pre-due buckets - around 38% - from accounts that were flagged and nudged before they ever became someone's commission opportunity.


What "moving on" actually looks like

Very few lenders fire every agency on day one, and they shouldn't. The pattern we see is a re-sequencing. Early and mid-bucket collections come in-house first, running on the platform, because that's where digital channels and prediction do most of the work and where agency commissions were least justified. Agencies get retained for what they're genuinely good at - late-stage NPA, legal-heavy pools, geographies where a field network takes years to build - and even those agencies increasingly work inside the lender's platform, so the data and the audit trail stay with the lender.


The end state isn't zero outsourcing. It's a collections operation where the lender owns the intelligence, the data and the borrower relationship, and buys labour only where labour is the actual constraint.


FrenzoFinserv exists to make that transition short: 4–6 weeks from first integration call to live portfolio, on your existing LMS, with your policies. If your agency contracts are coming up for renewal, that's usually the natural moment to run the comparison. Book a walkthrough and we'll model it on your buckets.