Powering one partner's multi-brand collections voice channel at scale
Oriserve became Credgenics' Generative AI voicebot partner, scaling from a single pilot engagement to 70+ bot builds across 16+ NBFC and bank brands — with the concurrent-call capacity to run India's highest-volume debt-collection campaigns.
About the partner
Credgenics is a collections and legal-recovery workflow platform that NBFCs and banks across India rely on to manage their loan servicing and debt-collection operations. Rather than build a voice AI capability in-house, Credgenics partners with specialized vendors to power the calling layer for its clients' campaigns — making Oriserve the voice engine behind Credgenics' outreach to borrowers across dozens of lending brands.
Partnership model: an aggregator/platform partnership — Credgenics brings the brand relationships, Oriserve powers the voice channel. Partner since September 2025. Use cases span pre-due reminders, Bucket X / Bucket 1 / Bucket 2 collections, write-off recovery, welcome calls, and EMI-bounce & e-mandate activation.
16+ NBFC and bank brands onboarded through the partnership, spanning consumer and small-business lending, vehicle and gold finance, credit cards and digital-first lenders.
The challenge
Being the voice layer for an aggregator means every new brand is a new deadline — at a volume few platforms are built for.
One partner, dozens of brands — every NBFC or bank onboarded needed its own bot, script and disposition logic, built and delivered fast.
Collections calling needs real scale — campaigns demand tens of thousands of calls an hour, sustained across the portfolio.
Language fragmented across geographies — borrowers span Hindi and English plus Tamil, Telugu, Kannada and Malayalam.
At this call volume, even small disposition-tagging gaps directly threatened invoicing accuracy and client trust.
Infrastructure had to flex with demand — concurrency, dialer throughput and server capacity growing as fast as the brand roster.
A path to independence, not dependency — Credgenics wanted to run campaigns and clone bots itself once the platform matured.
Beyond an off-the-shelf bot: the AI engineering we ran
Scale-grade infrastructure, backed by a service layer built for a fast-moving aggregator.
- Telephony and dialer scale built for collections. 300 concurrent connected calls and 50,000–80,000 calls/hour of dialer capacity, purpose-tuned for high-volume outbound debt collection.
- A cloning-first bot factory. Every new brand becomes a fast clone-and-adapt of a proven flow — Bucket X, pending charges, write-off, pre-due — instead of a rebuild.
- Rapid, hands-on UAT cycles. Dedicated PM support, on-site workshops at Credgenics' office, and same-day feedback turnarounds across a dozen bots running in parallel.
- Regional language engineering. Purpose-built Tamil, Telugu, Kannada and Malayalam bots, with Odia and Haryanvi in the pipeline.
- Disposition and billing accuracy engineering. Purpose-built disposition frameworks and backend reconciliation logic keep call outcomes — and invoicing — accurate at scale.
- A deliberate path to self-serve. An Agent Dashboard for cloning, testing and cURL-based configuration, plus structured KT sessions, so Credgenics can run campaigns independently.
Results, measured end-to-end
Beyond the headline numbers, the portfolio runs at a 25–30% connection (answer) rate with a 40–50 second average call handling time — steady even as monthly billed volume scaled roughly 17× since the September 2025 ramp-up, climbing consistently month over month and peaking at 17.5× in June 2026.
The solution we designed
Delivery model
Every new brand becomes a fast clone-and-adapt of a proven flow instead of a rebuild. Dedicated PM support, on-site workshops at Credgenics' office, and same-day feedback turnarounds keep a dozen bots running in parallel through rapid, hands-on UAT cycles.
Language coverage
Six languages are live today — Hindi, English, Tamil, Telugu, Kannada and Malayalam — purpose-built to match Credgenics' expanding brand footprint, with Odia and Haryanvi in the pipeline.
Infrastructure & scale
300 concurrent connected calls and 50,000–80,000 calls/hour of dialer capacity, purpose-tuned for high-volume outbound debt collection, with concurrency, throughput and server capacity scaled to keep pace with the growing brand roster.
Path to self-serve
Purpose-built disposition frameworks and backend reconciliation logic keep call outcomes — and invoicing — accurate at scale. An Agent Dashboard for cloning, testing and cURL-based configuration, plus structured KT sessions, lets Credgenics run campaigns independently rather than stay dependent on a permanent vendor bottleneck.
Why a generic solution would fail here
Brand-onboarding speed. A generic vendor would have struggled to keep pace with Credgenics' brand roster — every NBFC or bank onboarded needed its own bot, script and disposition logic, built and delivered fast.
Collections-grade scale. A generic vendor's infrastructure wouldn't sustain the tens of thousands of calls an hour a debt-collection portfolio demands, let alone 5–8 lakh calls a day.
Language coverage. A generic vendor would have struggled to deliver native-quality service across six languages.
Billing and disposition accuracy. At this call volume, even small disposition-tagging gaps directly threaten invoicing accuracy and client trust.
The takeaway
An aggregator platform lives or dies by how fast it can bring a new brand online and how much volume its calling layer can sustain. A generic vendor would have struggled to keep pace with Credgenics' brand roster, let alone deliver six languages and disposition-accurate billing at 5–8 lakh calls a day. By combining collections-grade telephony capacity with a cloning-first delivery model and a genuine path to self-serve, Oriserve grew alongside Credgenics — from a two-account pilot to the voice backbone behind 16+ lending brands, with monthly volume up roughly 17× along the way.
Have a workload like this?
We'll pull the closest live production analog and walk you through how the numbers were reached.
Powering one partner's multi-brand collections voice channel at scale
Oriserve became Credgenics' Generative AI voicebot partner, scaling from a single pilot engagement to 70+ bot builds across 16+ NBFC and bank brands — with the concurrent-call capacity to run India's highest-volume debt-collection campaigns.
About the partner
Credgenics is a collections and legal-recovery workflow platform that NBFCs and banks across India rely on to manage their loan servicing and debt-collection operations. Rather than build a voice AI capability in-house, Credgenics partners with specialized vendors to power the calling layer for its clients' campaigns — making Oriserve the voice engine behind Credgenics' outreach to borrowers across dozens of lending brands.
Partnership model: an aggregator/platform partnership — Credgenics brings the brand relationships, Oriserve powers the voice channel. Partner since September 2025. Use cases span pre-due reminders, Bucket X / Bucket 1 / Bucket 2 collections, write-off recovery, welcome calls, and EMI-bounce & e-mandate activation.
16+ NBFC and bank brands onboarded through the partnership, spanning consumer and small-business lending, vehicle and gold finance, credit cards and digital-first lenders.
The challenge
Being the voice layer for an aggregator means every new brand is a new deadline — at a volume few platforms are built for.
One partner, dozens of brands — every NBFC or bank onboarded needed its own bot, script and disposition logic, built and delivered fast.
Collections calling needs real scale — campaigns demand tens of thousands of calls an hour, sustained across the portfolio.
Language fragmented across geographies — borrowers span Hindi and English plus Tamil, Telugu, Kannada and Malayalam.
At this call volume, even small disposition-tagging gaps directly threatened invoicing accuracy and client trust.
Infrastructure had to flex with demand — concurrency, dialer throughput and server capacity growing as fast as the brand roster.
A path to independence, not dependency — Credgenics wanted to run campaigns and clone bots itself once the platform matured.
Beyond an off-the-shelf bot: the AI engineering we ran
Scale-grade infrastructure, backed by a service layer built for a fast-moving aggregator.
- Telephony and dialer scale built for collections. 300 concurrent connected calls and 50,000–80,000 calls/hour of dialer capacity, purpose-tuned for high-volume outbound debt collection.
- A cloning-first bot factory. Every new brand becomes a fast clone-and-adapt of a proven flow — Bucket X, pending charges, write-off, pre-due — instead of a rebuild.
- Rapid, hands-on UAT cycles. Dedicated PM support, on-site workshops at Credgenics' office, and same-day feedback turnarounds across a dozen bots running in parallel.
- Regional language engineering. Purpose-built Tamil, Telugu, Kannada and Malayalam bots, with Odia and Haryanvi in the pipeline.
- Disposition and billing accuracy engineering. Purpose-built disposition frameworks and backend reconciliation logic keep call outcomes — and invoicing — accurate at scale.
- A deliberate path to self-serve. An Agent Dashboard for cloning, testing and cURL-based configuration, plus structured KT sessions, so Credgenics can run campaigns independently.
Results, measured end-to-end
Beyond the headline numbers, the portfolio runs at a 25–30% connection (answer) rate with a 40–50 second average call handling time — steady even as monthly billed volume scaled roughly 17× since the September 2025 ramp-up, climbing consistently month over month and peaking at 17.5× in June 2026.
The solution we designed
Delivery model
Every new brand becomes a fast clone-and-adapt of a proven flow instead of a rebuild. Dedicated PM support, on-site workshops at Credgenics' office, and same-day feedback turnarounds keep a dozen bots running in parallel through rapid, hands-on UAT cycles.
Language coverage
Six languages are live today — Hindi, English, Tamil, Telugu, Kannada and Malayalam — purpose-built to match Credgenics' expanding brand footprint, with Odia and Haryanvi in the pipeline.
Infrastructure & scale
300 concurrent connected calls and 50,000–80,000 calls/hour of dialer capacity, purpose-tuned for high-volume outbound debt collection, with concurrency, throughput and server capacity scaled to keep pace with the growing brand roster.
Path to self-serve
Purpose-built disposition frameworks and backend reconciliation logic keep call outcomes — and invoicing — accurate at scale. An Agent Dashboard for cloning, testing and cURL-based configuration, plus structured KT sessions, lets Credgenics run campaigns independently rather than stay dependent on a permanent vendor bottleneck.
Why a generic solution would fail here
Brand-onboarding speed. A generic vendor would have struggled to keep pace with Credgenics' brand roster — every NBFC or bank onboarded needed its own bot, script and disposition logic, built and delivered fast.
Collections-grade scale. A generic vendor's infrastructure wouldn't sustain the tens of thousands of calls an hour a debt-collection portfolio demands, let alone 5–8 lakh calls a day.
Language coverage. A generic vendor would have struggled to deliver native-quality service across six languages.
Billing and disposition accuracy. At this call volume, even small disposition-tagging gaps directly threaten invoicing accuracy and client trust.
The takeaway
An aggregator platform lives or dies by how fast it can bring a new brand online and how much volume its calling layer can sustain. A generic vendor would have struggled to keep pace with Credgenics' brand roster, let alone deliver six languages and disposition-accurate billing at 5–8 lakh calls a day. By combining collections-grade telephony capacity with a cloning-first delivery model and a genuine path to self-serve, Oriserve grew alongside Credgenics — from a two-account pilot to the voice backbone behind 16+ lending brands, with monthly volume up roughly 17× along the way.
Have a workload like this?
We'll pull the closest live production analog and walk you through how the numbers were reached.