Detect, contain, correct: the quality loop

Quality in Workcell is a loop, not a checklist. You detect whether work meets spec, you contain what failed and decide what to do with it, and you correct the root cause so the same problem does not come back. Three record types carry that loop, each handing off to the next. This primer explains how they fit together; the how-to docs cover the clicks.

The three records

Detect: the inspection

An inspection is a recorded check against incoming material, in-process work, or a finished part. It does not invent its own checks. It is scored against an inspection template, which supplies the characteristics to check and their pass/fail criteria. The inspector records a result for each characteristic, then completes the inspection. If every result passed, it ends in Passed and the work moves on. If anything failed, it ends in Failed.

A failure is not the end of the story. A failed inspection acts on the thing it was checking: it holds the affected receiving lots, work order, or serial so the suspect material cannot move on, and it raises an NCR. That hold is the first act of containment.

Contain: the NCR

A nonconformance report (NCR) captures one specific failure: what was wrong, on what item, how much was affected, and what gets decided about it. Many NCRs are raised for you when an inspection fails; you can also raise one directly for a defect found any other way.

Working an NCR has two jobs. First you record the root cause so you understand why it happened. Then you set a disposition: the decision for the affected material, such as use as is, rework, scrap, or return. The disposition is what containment is for. It answers "what do we do with the parts we already have," and an NCR cannot be closed until that decision is recorded. Containment deals with the batch in front of you. It does not, on its own, stop the next batch from failing the same way.

Correct: the CAPA

That is the job of a CAPA, a corrective and preventive action. When a nonconformance is recurring or significant, containing this batch is not enough. You need to fix the cause so it stops happening. A CAPA usually comes from an NCR and links back to it, so the lasting fix is always traceable to the failure that prompted it.

A CAPA runs through the 8D disciplines, D0 through D8, one step at a time: forming a team, describing the problem, containing it, finding the root cause, putting corrective actions in place, validating them, and preventing recurrence. Completing the prevent-recurrence step moves the CAPA to Pending verification, and closing it records that the fix is in place.

Why it is a loop, not one big form

Each record answers a different question and has a different owner. The inspection asks "does this conform?" The NCR asks "what do we do with what failed?" The CAPA asks "how do we stop it happening again?" Splitting them into linked records means a disposition never gets confused with a root-cause fix, and you can always trace a CAPA back through the NCR to the inspection that first caught the problem.

It is also why quality is a loop rather than a line. A CAPA that prevents recurrence feeds back into the work, so the next inspection is more likely to pass. Detect, contain, correct, and the loop closes on itself: the point of finding a problem is to make finding it again unnecessary.

Ready to walk it? Start with Run an inspection, then Raise an NCR and Manage a CAPA.