Three Days Down, Two Broken Promises
You are the floor supervisor on an inbound broadband support process in a Jaipur BPO. Sixty agents take calls in Hindi and English. The contract sets an average handling time target of 420 seconds, a first-call-resolution target of 78% and a customer satisfaction target of 85%. An agent, Ritu, has just transferred a call to you after nine minutes. The customer's home internet has been down for three days. Two earlier agents promised a technician within 24 hours; no technician came and no callback happened. The customer has been charged ₹1,200 for the month, is shouting, has threatened to post the recording on social media, and is demanding three months free. Your authority is a credit of up to ₹500 and a priority technician visit within 48 hours. Quality audit penalises agents for long calls but also penalises any promise that is not kept. Ritu's quality score is already 71% against an 80% threshold.
1. You pick up the call. Lay out, in order, what you say and do, and what you must not say.
Read the worked answer
Order matters, because trust is rebuilt before it is spent.
- Identify yourself with name and role - supervisor, not another agent - and apologise once, specifically: three days without service and two missed appointments is not what was promised. One clear apology, not five vague ones.
- Verify identity before discussing the account. Confirm name, registered mobile or account number. Data privacy means you do not confirm the address, the plan or the billing history to an unverified caller, however angry they are.
- Let them finish without interruption for the first minute and take notes on dates and commitments.
- State plainly what went wrong: two commitments were logged and not kept. Own it as the company; do not name or blame the earlier agents.
- Offer exactly what you have authority for: a ₹500 goodwill credit, described as goodwill and not as a refund of the ₹1,200 - do not let the customer infer that the billing dispute is settled - and a priority technician with a booked slot within 48 hours plus a ticket number.
- Say what happens next and when: an SMS with the ticket number and technician window, a callback after the visit, and escalation to a senior if it fails again.
Must not say: anything about three months free; any amount above ₹500; a specific visit time that scheduling has not confirmed; 'calm down'; 'this is not my department'; 'I know how you feel' as a substitute for action. And never hang up first.
2. Ritu wanted to disconnect the call; the AHT clock was already at 700 seconds. Decide what happens to the call and to Ritu, using the metrics.
Read the worked answer
Read the metric correctly first. A 420-second average handling time is an average, not a per-call cap. Judging one call against an average is a statistical error, and a supervisor who does it teaches agents to end calls rather than resolve them.
Cost out the failure. A repeat call, an escalation, a public social-media post and possibly a lost customer cost far more than the 280 seconds over target. First-call resolution at 78% and CSAT at 85% exist precisely to stop teams from optimising AHT in isolation.
Decision on the call. It continues until the customer has a ticket number and a booked slot. Nothing is gained by ending it earlier.
Decision on Ritu. The transfer at nine minutes was correct, not a failure: she had no authority for the credit or the priority slot, and holding the customer without authority would have made things worse. So no penalty for the long call and no quality defect for the escalation - log it as an authority-limit escalation. Coach her on earlier escalation, around the five-to-six minute mark, once she knows the promise has been broken twice.
The 71% score is a separate matter. Address it with a coaching plan and a review of which parameters are failing, not by defending the process in a live call. And state the boundary clearly: disconnecting a customer is a disciplinary matter regardless of what the AHT clock says.
3. Identify the root cause and propose one process change with a measure that would show it worked.
Read the worked answer
Look past 'the agent was careless'. Three structural causes explain the failure better than individual blame.
- Agents can record a promise without checking technician availability, so the commitment is invented at the moment of pressure.
- No callback task is created when a promise is made, so nobody owns the follow-up.
- The next agent cannot see that the customer has already been promised twice, so the same promise is repeated.
Add the incentive problem: AHT rewards closing the call, not keeping the promise.
One process change. Make the promise a record. The CRM requires the agent to select an available technician slot before the commitment can be read to the customer; that selection automatically creates a callback task, and the next call screen shows 'prior commitments: 2 missed'. This is configuration and scripting, not headcount.
Measures. Repeat contacts within 72 hours for the same connection should fall from the current baseline; a weekly audit of 30 random calls should score 'promise kept within the stated window'; and technician no-show rate should be reported daily, since that is a different failure with a different owner.
Watch for the side effect. Expect AHT to rise slightly in the first month. Treat the target as a band and watch for agents gaming it by transferring, otherwise the fix quietly moves the problem downstream. Review at six weeks; if repeat contacts have not moved, the constraint is technician capacity, not the promise process.