Agent QA Scorecard Template for Collections Call Centres
An agent QA scorecard for collections scores each call across weighted categories — compliance, communication, process and resolution — usually out of 100. For Indian collections, RBI recovery-conduct rules (calls only between 8:00 a.m. and 7:00 p.m., no harassment) should sit as an auto-fail compliance gate, not just a scored line.
This is an operational explainer, not legal advice. For binding obligations, refer to the official RBI circular and consult your compliance team.
In collections, quality assurance (QA) is where compliance and recovery meet. A good scorecard does two jobs at once: it protects the lender from conduct breaches that carry regulatory and reputational risk, and it coaches agents toward better promise-to-pay outcomes. This guide gives you a ready-to-use template, with the Reserve Bank of India's recovery-conduct rules built in as hard gates.
What Is an Agent QA Scorecard for Collections?
A QA scorecard is a structured evaluation form. Reviewers — human or AI — listen to a call and score it against defined categories, each carrying a weight so the total lands on 100. It turns a subjective "that was a good call" into a consistent, comparable, coachable number.
For collections specifically, the scorecard must encode conduct rules. The RBI's circular on recovery agents — RBI/2022-23/108, dated August 12, 2022 — directs that regulated entities and their agents must not harass or intimidate borrowers, and must not call before 8:00 a.m. or after 7:00 p.m. for recovery of overdue loans (RBI, 2022). These are not "points off" items — a breach should fail the entire call.
The Ready-to-Use Collections QA Scorecard Template
Copy this scorecard and tune the weights to your priorities. The weights sum to 100; the compliance category also carries auto-fail rules that override the score.
Category | Weight | What it checks |
|---|---|---|
Compliance & Conduct | 30% | Call within 8:00 a.m.–7:00 p.m.; no harassment, intimidation or threats; correct identification and disclosure; no false representation |
Opening & Verification | 10% | Proper greeting, right-party contact confirmed, purpose stated clearly |
Communication & Empathy | 20% | Respectful tone, clear language, active listening, no aggressive pressure |
Process Adherence | 15% | Accurate account details, correct dispute handling, approved script followed |
Resolution & Negotiation | 15% | Payment options offered, objection handling, realistic promise-to-pay (PTP) secured |
Documentation & Disposition | 10% | Accurate call notes, correct disposition code, follow-up scheduled |
Total | 100% | — |
Auto-fail rules (override the total, score = 0):
- Any call placed outside the 8:00 a.m.–7:00 p.m. window for an overdue-loan recovery contact
- Any harassment, intimidation, threat or public-humiliation language
- Any false or misleading representation to the borrower
These mirror the conduct the RBI circular directs agents to avoid (RBI, 2022), reinforced by the long-standing Fair Practices Code for Lenders.
Why Should Compliance Be an Auto-Fail, Not Just a Weighted Line?
Because the lender — not the agency — carries the liability. The RBI is explicit that ultimate responsibility for outsourced recovery vests with the regulated entity (RBI, 2022). A call that recovers the full amount but harasses the borrower or lands at 7:45 p.m. is a failure regardless of the recovery, because it exposes the lender to grievance and regulatory risk. Making conduct an auto-fail forces that priority into every review. For the full rule set, see our breakdown of ensuring voice AI compliance with RBI guidelines.
How Do You Score and Weight Each Call?
Score each category on its own scale, convert to its weighted share, and sum — unless an auto-fail rule triggers, in which case the call scores zero and routes straight to coaching. A worked example:
Category | Weight | Raw score (0–100) | Weighted contribution |
|---|---|---|---|
Compliance & Conduct | 30% | 100 | 30.0 |
Opening & Verification | 10% | 90 | 9.0 |
Communication & Empathy | 20% | 80 | 16.0 |
Process Adherence | 15% | 85 | 12.75 |
Resolution & Negotiation | 15% | 70 | 10.5 |
Documentation & Disposition | 10% | 90 | 9.0 |
Final QA score | 100% | — | 87.25 |
Set a pass threshold (commonly 80–85) and a coaching band below it. Track scores by agent, team and category over time so you can see whether the floor is improving on empathy, negotiation or documentation — and whether conduct auto-fails are trending to zero. Our guide on AI call monitoring and agent performance in banking covers the metrics that matter.
How Do You Move From Sampling to 100% QA Coverage?
Manual QA typically reviews 2–5% of calls — a small, potentially unrepresentative sample. The gap matters most in collections, where a single out-of-window or harassing call can trigger a complaint. AI-based conversation intelligence lets you score every call against the same scorecard, so compliance auto-fails are caught across the full volume rather than by luck of the sample. This pairs naturally with the recovery use cases in our complete India collections playbook and voice AI use cases for NBFC collections.
How AI Helps Run the Collections QA Scorecard
YuCI, YuVerse's conversation intelligence platform, applies your scorecard to 100% of calls instead of a small sample. It flags conduct breaches automatically — calls outside the 8:00 a.m.–7:00 p.m. window, aggressive or threatening language, missing disclosures — as auto-fails, and scores communication, negotiation and documentation consistently across agents. Because every call is transcribed and timestamped, supervisors get an objective audit trail and a coaching queue that points to the exact moments to fix, so quality and compliance improve together across the whole floor.
FAQ
What is an agent QA scorecard in collections? It is a weighted evaluation form used to score collections calls out of 100 across categories such as compliance, communication, process adherence and resolution. It standardises quality reviews and drives agent coaching.
What RBI rules must a collections QA scorecard enforce? The August 12, 2022 RBI circular directs that recovery agents must not harass or intimidate borrowers and must not call before 8:00 a.m. or after 7:00 p.m. for overdue-loan recovery (RBI, 2022). These belong as auto-fail gates in the scorecard.
How should compliance be weighted? Compliance should carry a heavy weight (around 30%) and, critically, act as an auto-fail: any conduct or calling-window breach fails the whole call, because the lender bears ultimate responsibility for recovery conduct (RBI, 2022).
What categories should the scorecard include? A practical set is Compliance & Conduct, Opening & Verification, Communication & Empathy, Process Adherence, Resolution & Negotiation, and Documentation & Disposition, with weights summing to 100.
What pass score should we set? Most teams set a pass threshold of 80–85, with a coaching band below it. Any auto-fail scores zero regardless of the rest of the call and routes straight to coaching and, where needed, remediation.
Can AI score every collections call? Yes. Conversation intelligence platforms like YuCI apply the scorecard to 100% of calls, flag compliance breaches automatically, and produce a timestamped audit trail — far beyond the 2–5% that manual QA typically samples.
Conclusion
A collections QA scorecard has to serve two masters — recovery and conduct — and the way to reconcile them is to make RBI compliance an auto-fail gate over a weighted quality score. Build the template around clear categories, weight compliance heavily, and move from sampling to full-coverage AI review so no out-of-window or harassing call slips through. That is how collections floors improve outcomes and stay inside the rules at the same time.
Want to apply your QA scorecard to every collections call? Talk to the YuVerse team.
References
- RBI — Outsourcing of Financial Services: Responsibilities of regulated entities employing Recovery Agents (RBI/2022-23/108, August 12, 2022) — https://www.rbi.org.in/Scripts/NotificationUser.aspx?Id=12378&Mode=0
- RBI — Guidelines on Fair Practices Code for Lenders (May 5, 2003) — https://www.rbi.org.in/Scripts/NotificationUser.aspx?Id=1172&Mode=0
