One straightforward example is a part dimension that is outside the tolerance defined on the drawing or work instruction.
For example, if a hole is specified as 10.00 mm plus or minus 0.05 mm, and inspection records 10.12 mm, that is a non-conformity because the part does not meet the approved requirement.
Other common examples include:
- Using the wrong material lot or an unapproved substitute
- Missing or incomplete inspection records
- Assembly performed to an obsolete revision of the work instruction
- A calibration status that is expired at the time a gage was used
- Damage, contamination, or cosmetic defects when appearance requirements apply
- Process parameters run outside the approved range
Not every non-conformity has the same consequence. Some lead to rework, some to scrap, some to use-as-is or repair requests where the quality system allows that path, and some may trigger broader corrective action if the issue is systemic. The correct disposition depends on the requirement violated, product criticality, customer or regulatory constraints, and the organization’s approval workflow.
In regulated manufacturing, the key point is not just that something is wrong, but that the condition is identified, documented, contained, and traceable. A missed record, wrong revision, or broken genealogy link can be a non-conformity even when the physical part looks acceptable.
In brownfield environments, these events are often detected across multiple systems such as MES, ERP, QMS, paper travelers, and inspection software. That means classification and response quality depend heavily on integration quality, master data alignment, and disciplined change control. Replacing all systems is usually not the practical answer in long-lifecycle regulated operations because validation effort, downtime risk, and traceability disruption can outweigh the benefit.