CASE 04 · CASE ATLAS · LEVEL B WORKSHOP SCENARIO

Repeated bolted-joint issues

SITUATION

The same bolted-joint torque non-conformity keeps recurring on a tower platform across multiple sites. Each site has opened its own corrective action, but no one has compared them, so the investigations restart from zero each time.

EVIDENCE

Closed and open corrective-action records across sites, each with a described symptom, root cause and containment action, plus the current investigation's symptom description.

AI APPROACH

Retrieval clusters prior corrective actions by symptom similarity to the current investigation, surfacing candidate matches with their stated root cause and containment, for the investigator to judge relevance.

KEY CONTROLS

Clustering is presented as candidate similarity, not confirmed recurrence; the investigator must independently verify that a retrieved prior case shares the actual failure mechanism, not just similar wording.

VALIDATION MEASURES

Measure whether the retrieval surfaces the corrective actions a manual cross-site search would find, and how often a surfaced match turns out to be a false lead the investigator has to discard.

HUMAN DECISION

The investigator decides which prior corrective actions are genuinely relevant and whether the recurrence justifies escalating to a systemic root cause across sites.

DISCUSSION QUESTIONS

  1. What is the cost today of not knowing this failure has recurred at three sites?
  2. How would a wrongly clustered 'similar' case waste investigation time, and how would you catch it?
  3. Who owns closing the loop if the retrieval reveals a systemic cause beyond any single site's corrective action?

TRY IT LIVE

Run this scenario's approach through Quality Evidence Graph — What connects this decision to its evidence, controls, and lessons learned?

Open Quality Evidence Graph →