The checklist that was completed during a breakdown
During a shop-floor interview a press operator tells the auditor that first-off parts are checked "when we remember, mostly at the start of a run". The line supervisor, who joins the conversation, says checks are done every two hours and produces the day's checklist: eight entries, all ticks in the same pen, including the 10:00 and 12:00 slots. The maintenance log shows the press was down from 09:30 to 13:00 for a hydraulic repair. The operator looks at the floor. The supervisor says the checklist "must have been filled in later, to catch up".
1. What are your next steps, and what evidence do you gather before forming a conclusion?
Read the worked answer
Do not resolve it in the moment and do not accuse anyone. Record verbatim what each person said, with times.
Triangulate from independent sources:
- Maintenance log and hydraulic repair ticket - already contradicts two of the eight claimed check times.
- Production count and machine or ERP timestamps over the eight claimed intervals: was product made at 10:00?
- First-off approval and inspection records for the run.
- Checklists for the previous 10 working days: identical ticks, single pen and suspiciously uniform entries indicate retrospective completion as a pattern rather than a one-off.
- Interview two other operators and the quality inspector separately, without the supervisor present.
- Ask how the form is controlled and who reviews and signs it (7.5.2).
In the finding, describe facts, not motives: "checklist entries record checks at 10:00 and 12:00 on a press recorded as under repair from 09:30 to 13:00". That is defensible; "the supervisor falsified records" is not, unless you can evidence intent.
2. Is this a nonconformity, and against which clause or clauses? How would you avoid over-issuing?
Read the worked answer
Primary nonconformity - 8.5.1. Monitoring and measurement at defined stages is not implemented as specified: entries were recorded for periods when production could not have occurred. Evidence: the checklist plus the maintenance record.
Upstream question. The operator's "when we remember" may indicate the check frequency is undefined, under-resourced, or incompatible with the run length. If so, that is part of the same condition, not a separate finding.
Possible second finding - 7.5.3 / 8.6. Documented information used as evidence of conformity is not accurate, which undermines release decisions.
Avoid duplication. One clear nonconformity carrying all the evidence is stronger than three that restate the same condition; only issue separately if the record-integrity failure extends to processes outside this press.
Severity. Lean major if product was released on the basis of those records and the practice is widespread; minor if it is confined to this line and conformity is verified elsewhere - pending the sampling in task 1.
3. How do you handle the supervisor's admission and the operator's discomfort, both in the room and in your report?
Read the worked answer
In the room. Treat the admission as evidence, not as a confession to be pursued. Thank them, record it verbatim, and ask one neutral clarifier - "when catching up happens, who checks the entry?" - without leading or signalling blame. Do not comment on discipline, and do not promise confidentiality you cannot keep.
Protect the operator. The operator is the more vulnerable party; keep their name out of notes and report, identify by role, and avoid a follow-up interview in front of the supervisor. Body language is not evidence - note it, do not build on it.
Escalation. Where records appear to have been created retrospectively, this is a management-integrity issue: raise it with the auditee's top management at the closing meeting as a factual finding and a risk to product release, not as an accusation against a named individual. Follow your own audit programme's procedure for escalating suspected falsification.
State the limitation. One day's records and three interviews cannot establish how widespread the practice is; say so in the report and recommend a wider sample in the next audit.