After the audit, everyone knows who takes the next step.

An audit found an outdated component inspection instruction. Updating the file is only part of the work. The organisation must determine what changed, which production was affected, who can approve the new instruction and how the team will check that the original problem stopped recurring. Otherwise the next audit finds the same weakness under another document name.

This guide is for quality managers, production owners and manufacturing leaders who need a manageable path from a finding to a supported decision. The workflow and examples are our proposed operating method. Example names, quantities and timings are hypothetical. They are not customer results or universal requirements of a certification scheme.

Start with the requirement that actually applies

Manufacturing sites can face internal procedures, customer specifications, contract conditions, legal requirements and certification standards at the same time. Keep their sources separate. A customer may specify an inspection frequency that is stricter than the internal instruction. An auditor may assess a different scope from the customer visiting next month. One combined list without source and version hides those differences.

As of 2 October 2026, ISO publishes ISO 9001:2026 as the current edition. Use a lawful copy and assess the transition with the certification body. Historical evidence and IATF 16949:2016 references should retain their original context. Automotive suppliers also check the current requirements of their actual OEM customers. Packaging producers using BRCGS check the applicable Packaging Materials scope and issue rather than borrowing a food-manufacturing checklist.

For each finding, record the source document, version, applicable process and expected condition. Then record what was observed and the evidence supporting it. “Instruction outdated” is too vague. A useful hypothetical finding states that the instruction at Workstation 12 uses revision B while the approved control plan and customer change require revision C for Product X.

Assess the affected production before changing the document

The first decision concerns impact. Determine when the wrong instruction could have been used, which products were manufactured and which records show what inspection actually occurred. Do not assume that a document discrepancy proves defective product. Do not assume that a complete production record proves conformity either. The quality and process owners assess the available evidence against the applicable release criteria.

In the worked example, the old instruction omits an additional inspection after a tooling change. The team identifies the tooling-change date and relevant production lots. It checks whether another controlled record captured the inspection. Where evidence is insufficient, the authorised owner decides on containment and further assessment. The decision, reason and affected population are retained.

Pulsar GRC can organise the finding, action, owners and supporting documents. It does not perform production measurement or decide whether a component can be released. Those decisions require the organisation’s technical competence, process data and authority. Keep the operational decision with the appropriate owner and connect its record to the audit case.

Separate immediate correction from cause removal

Replacing the workstation copy with the approved instruction corrects the visible situation. It does not explain why the copy remained obsolete. Examine the document-change workflow: who approved revision C, how affected stations were identified, how printed copies were replaced and how completion was verified. Check actual records before selecting a cause.

Possible explanations include an incomplete station list, a missed night shift, an uncontrolled local printout or a change that never entered the approved distribution route. Each leads to a different action. More training will not repair an incomplete station list. A new distribution email will not remove old printed copies. Choose the action that addresses the supported mechanism.

Record uncertainty honestly. If the team cannot establish whether the old copy was used, state the limitation and assess the consequences. Do not write a confident root cause merely because the register requires text. A further investigation task with an owner is a legitimate next step when the available evidence is incomplete.

Give every action an acceptance criterion

An action should state the result, responsible person, due date and evidence required for acceptance. “Update instruction” becomes: revise the affected inspection instruction, approve it through the agreed route, replace controlled workstation copies and demonstrate that the old version is no longer available for routine use. The organisation chooses details appropriate to its process.

The implementation evidence might include the approved revision, change assessment, station distribution list and observed replacement check. Keep the individual records connected so that the reviewer can see how the change reached production. A signature on the revised document proves approval; it does not prove replacement at every affected station.

Where the change affects competence, involve the people who actually perform or supervise the task. CrewShift can support training, acknowledgements and competency-related work within the current offer. A document acknowledgement alone does not establish the ability to execute a changed inspection. Use observation or a relevant exercise if that ability needs verification.

Design the effectiveness review before closing the case

Choose the review criterion before the results are known. For the hypothetical instruction case, the team might check all affected stations across both shifts, review the next relevant tooling changes and confirm that the distribution route includes each affected role. These are example criteria, not a mandated universal sample.

Separate implementation from effectiveness. Implementation asks whether the new instruction and distribution change were introduced. Effectiveness asks whether subsequent changes reach the right stations and whether the process uses the correct revision. A newly printed instruction can pass implementation while the next engineering change still follows the failed route.

Record the opportunities available for observation. A frequent tooling change may produce several meaningful checks quickly. An infrequent product change may need a longer observation period or a controlled exercise. If the planned event has not occurred, mark effectiveness pending. A due date does not make an untested control effective.

Include exceptions in the review. Check absence of the usual approver, an urgent customer change or a workstation added after the first distribution list. These are the conditions under which a seemingly sound process can fail. If the test reveals another bypass, revise the analysis and preserve the first result rather than deleting inconvenient evidence.

Keep production systems and the GRC record connected

The GRC case should point to the authoritative production evidence without becoming a substitute for the system that creates it. Inspection results, SPC data, laboratory records and manufacturing traceability may remain in their specialist systems. Identify the relevant record, version, period and access method. Copy only what the review needs and what the organisation is authorised to use.

Avoid a second uncontrolled version of an operational instruction. If the approved source changes, record the change and update the reference used by the case. If an export is needed for an auditor, label the scope and export date. The organisation should be able to explain whether the shared file is the original, a controlled copy or a limited extract.

Sensitive records may contain customer designs, employee information or commercial terms. Limit the audit package to the agreed scope and recipients. A complete evidence trail does not require giving every reviewer access to the whole site archive. Record the sharing decision and any justified redactions so the recipient understands what was supplied.

Use one pilot to test the operating method

Choose one recent finding that involves a changed instruction and several owners. Follow it from the source requirement to impact assessment, cause analysis, action and effectiveness decision. Measure the time needed for a colleague to retrieve a complete case. Count missing owners, rejected evidence and decisions that still need private email context.

A useful pilot also tests handover. Ask a colleague who did not create the case to explain what happened, which production was affected and why the action was closed or remains open. If the colleague cannot answer, identify the missing relationship or record. Repair the workflow before importing hundreds of findings.

Compare the pilot with your own baseline. Shorter retrieval time is useful only if the answer remains complete and the maintenance effort is manageable. Record recurring work required to keep documents and decisions current. A new system that creates a parallel register may move effort rather than reduce it.

Make the closure decision explicit

The authorised reviewer closes the case only when the agreed criteria are supported, or records a justified alternative decision under the organisation’s procedure. The record identifies the evidence reviewed, result, residual risk and any continuing monitoring. The person who implemented the change may contribute evidence while a different role reviews it where the procedure requires that separation.

If the problem recurs, retain the previous closure and link the new finding. Determine whether the original cause was misunderstood, the control degraded or a later change created another route. Recurrence is evidence about the process. Hiding it in a new unconnected task makes the dashboard cleaner while leaving the mechanism untouched.

Pulsar GRC supports audits, actions and documentation. It does not replace an SPC system, production inspections, a licensed standard or an auditor’s assessment, and it does not guarantee certification. AI-supported structuring produces proposals that a person checks against the organisation’s sources and actual process.

Check that a later change does not revive the old gap

After closure, connect the action to the process that manages future instruction changes. A new workstation, product variant or customer requirement may invalidate the distribution list used in the original case. Define who recognises that event and which record is reassessed. The instruction-control method needs to survive the next change rather than only the current audit.

For a hypothetical added station, the process owner checks approved documents, required inspections and role preparation before routine use. The GRC record can retain the assessment and action references. Production systems still perform the technical checks. If the new station is outside the old case scope, record that difference and open the necessary work rather than extending the previous closure without evidence.

This review trigger is our operating recommendation. It gives the next responsible person a clear reason to revisit a closed case when the manufacturing context changes.

Start with the manufacturing workflow demo and a representative case using suitable sample data. Check the current features and offer. The current trial is 14 days and requires a payment method. The decision should rest on a working evidence trail that your team can maintain during normal production.