Voice AI for Loan Collections: How Rootle Powers Oroboro’s Borrower Calls

Three professionals in a meeting: a man in blue writes on a pad while two colleagues smile near a laptop on the table.

Increase in Hiring Quality

Time saved by recruiters

Location

Pan India

Industry

NBFC

Oroboro

Oroboro is a financial inclusion platform built around a simple idea, the one baked into its name: what goes around, comes around. It's owned by Transwarranty Finance Limited, an RBI registered NBFC that's been serving Indian borrowers since 1994, alongside its SEBI registered broking arm, Vertex Securities.

The platform runs three lines under one roof: small ticket pocket loans, insurance across life, health, auto and home, and wealth management for investors who want more than a savings account. The loans book is where most of the daily calling happens, and that's where this story starts.

1994 Transwarranty Finance founded, RBI registered NBFC

₹25K Typical pocket loan ticket size, repaid over 9 EMIs

3 Product lines under one platform: Loans, Insurance, Wealth

The Opportunity

More Borrowers, More Calls

A ₹25,000 loan means frequent EMIs and a large number of borrowers to stay in touch with regularly. Many of Oroboro's customers don't have a fixed salary date each month, so a well timed reminder can make a real difference to on time repayment. As the loan book grows, that adds up to a lot of conversations, each one worth getting right.

There's a compliance layer built into every one of those calls, too. NBFCs operate under the RBI's Fair Practices Code, which sets clear rules for when a borrower can be called and how a missed payment gets discussed. Oroboro wanted a voice AI system that could work confidently within these guidelines at scale, bringing the same care a trained agent would to every single call.

Repayment reminders go out on a steady cycle, timed to each of the nine EMIs across the loan.

A promise to pay is only useful once it's tracked and followed up at the right time, across thousands of borrowers.

Co-borrowers and guarantors are already part of every loan agreement. Knowing when to bring them into a call matters.

Hindi is the language most calls run in. Borrowers often mix in English mid sentence, and the call needs to follow along.

Rootle's Consultative Approach

Built Around Oroboro's Loan Agreements

Most collections technology starts from a recovery target and works backward. Rootle starts from the relationship. Before writing a single line of calling logic, we sat down with Oroboro to understand something most vendors never ask: what does a borrower actually need to feel respected on this call, even when they're behind on a payment? That conversation shaped everything that followed, and it's the reason Oroboro's loan agreements, not a generic collections template, define exactly who gets called, when, and what happens after a missed payment.

Start with the loan agreement itself

Every escalation trigger, every guarantor clause, and every dispute category came straight out of Oroboro's own documentation, mapped out before any calling logic was written.

Build around the Fair Practices Code from day one

Call windows, tone, and language were all built against RBI guidelines from the start, so every call already sits comfortably within what's permitted.

Plan for the full range of calls that come in

Disputes, wrong numbers, and unreachable borrowers make up a large share of collections calls. We mapped these carefully alongside the standard reminder flow.

Let the conversation set the language

Rootle listens for what the borrower actually speaks and adapts as the call happens.

Call Flow One

Reminder to Promise to Pay

Most Oroboro calls are routine, and routine is exactly where Rootle needs to be dependable. Here's how a standard EMI reminder plays out, from the first ring to the follow up that closes the loop.

Voice Call

Pre due date reminder

A day or two before the EMI is due, Rootle places a short, friendly call confirming the amount and the date.

Voice Call

Missed EMI follow up

If the date passes with no payment, a follow up call goes out within a set window to check in.

Voice Call

Promise to pay captured

If the borrower commits to a date, Rootle logs the exact promise, amount and reason if one's given, straight into the loan record.

WhatsApp & SMS

Payment link shared

Right after the call, a payment link goes out over WhatsApp and SMS, so paying takes one tap instead of a callback.

WhatsApp & SMS

Promise date reminder

A quick reminder and the payment link land again the day before the promised date, to help the borrower stay on track.

Loop closed or routed onward

When payment comes in, a receipt is emailed automatically and the loop closes. If the promise isn't kept, the call moves into the escalation flow.

Voice, for the conversation

WhatsApp & SMS, for the payment link

Email, for the receipt

Call Flow Two

Handling Escalations and Disputes

Some borrowers raise questions about what's owed. Others go quiet after a broken promise or two. Rootle recognizes each situation as it happens and routes the call to the right next step.

A payment is overdue and the call is underway

Escalation Path

When a promise isn't kept

A first missed promise gets a gentle, understanding follow up call.
A second missed promise, or overdue days crossing the set threshold, moves the account to a human collections agent.
The agent picks up with full context already attached, including prior promises, call history, and days overdue.

Dispute Path

When a borrower disputes the amount

If a borrower questions the amount, an interest charge, or says a payment's already been made, Rootle stops there.
The call transfers straight to a trained human agent.
The dispute reason and a summary of the conversation move with the call, so the borrower doesn't have to repeat their concerns over and over.

Call Flow Three

Involving the Co-Borrower or Guarantor

Every Oroboro loan agreement names a co-borrower or guarantor. Rootle reaches out to them only after the primary borrower has genuinely been unreachable, and stays within what the agreement allows.

Primary borrower unreachable

A set number of unanswered attempts across the reminder cycle is what triggers this outreach.

Consent already on file

Rootle only calls the co-borrower or guarantor named in the signed agreement, and only discusses what that agreement covers.

A clear, factual update

The conversation covers what's outstanding, since when, and how to reach the borrower or make a payment directly.

Built Into Every Call

Adapting to the Borrower's Language

Most of Oroboro's borrowers speak Hindi day to day, so that's where every call starts by default. Borrowers often move between Hindi and English within the same sentence, and Rootle follows along naturally as the conversation happens.

Hindi Default opening language for every outbound call

Hindi + English Follows natural code switching mid conversation

Adaptive Shifts fully if the borrower responds in another language

The call adjusts to how each borrower actually speaks, in the moment, without needing them to ask.

By the Numbers

Calling Capacity, Built for Scale

A few numbers that show what Rootle brings to Oroboro's collections calling on a typical day.

72% Average connect rate on first attempt, well above typical manual dialing

1,000+ Collection calls handled per day, without added headcount

<10 sec Average time to connect once a call is placed

24/7 Coverage for reminders and promise to pay follow ups

The Commitment

What Rootle Is Built to Deliver

These are the standards Rootle is built to meet from day one for Oroboro's collections calling.

Of calls placed within the RBI permitted calling window, every time

Of promise to pay commitments logged and followed up automatically

Disputes handled by the AI alone. Every one transfers to a trained agent

Languages handled fluently within a single call, Hindi and English, without a menu

Why Rootle

Built for Lenders Who Can't Cut Corners on Compliance

Oroboro isn't the only NBFC balancing collections volume against the Fair Practices Code. It's a familiar spot for regulated lenders once the loan book starts growing faster than the calling team.

Compliance built into every call

RBI calling windows and Fair Practices Code language are part of the calling logic itself.

Language that follows the borrower

Hindi first, English when needed, and the mix in between, all handled within a single conversation.

Every call feeds one record

Promises, disputes, and guarantor outreach all land in the same loan record, so the next call, human or AI, starts with the full picture already there.

A small ticket loan book runs on thousands of small conversations. Rootle's built to handle every one of them well, in the borrower's own language, and fully within the rules that protect them.