Decision Lab
A managerial laboratory, not a quiz. Read the evidence, inspect the diagnostic constructs, weigh the alternatives and commit to a decision you can defend.
Apex Auto Components Ltd. is a 450-employee tier-1 supplier of precision machined parts to two passenger-vehicle OEMs. The board has approved, in principle, an AI-assisted visual quality inspection system on the two main machining lines. Over the last two quarters productivity has fallen 8%, in-process rejection has risen from 3.2% to 5.1% and unplanned absenteeism stands at 6%. Operators believe the AI system is a monitoring tool that will be used to reduce headcount; supervisors believe operators are careless. The plant head must decide how to proceed within this quarter.
Employees
450
Productivity change
-8% over two quarters
Rejection rate
3.2% -> 5.1%
Unplanned absenteeism
6%
Proposal on the table
AI-assisted quality inspection
This is a fictional composite case written for teaching. Treat the diagnostic scores as evidence to interpret, never as an algorithmic verdict about any individual.
Psychological Safety
Management Trust
Job Autonomy
Role Clarity
Team Cohesion
Change Readiness
Digital Self-efficacy
Job Engagement
Below 2.6 signals a construct needing direct intervention; 2.6–3.2 is fragile; above 3.2 is a relative strength.
Alternatives and analysis
Tighten output targets, increase supervisory checking on both lines and move a larger share of pay to individual incentives.
Draws on Theory X assumptions and operant reinforcement through contingent pay. It can lift short-term output, but with psychological safety at 2.4 and trust at 2.5 it risks suppressing error reporting further and raising counterproductive behaviour.
Pilot the AI inspection on one line with operator participation in rule-setting, structured training, supervisor coaching in consideration behaviours, and cross-functional quality teams.
Rests on Lewin's unfreezing through participation, self-determination theory (autonomy, competence, relatedness), Edmondson's psychological safety and Kotter's guiding coalition. It is slower to show output gains but works directly on the diagnosed deficits.
Postpone the AI investment and recruit additional quality inspectors to contain rejection through manual checking.
A structural buffering response consistent with contingency thinking. It protects the customer in the short term, but adds recurring cost and treats rejection as an inspection problem rather than a behaviour and capability problem.
Roll out AI inspection across both lines immediately, then redesign jobs and reassign operators after the system stabilises.
A technology-first, change-by-fait-accompli approach. Speed can create momentum, but with change readiness at 2.6 and digital self-efficacy at 2.8 resistance, absenteeism and quality escapes are likely to spike during roll-out.
Set the weight of each criterion (1–5), then score every alternative (1–5). The matrix is a thinking aid, not a verdict.
| Criterion | Weight | A | B | C | D |
|---|---|---|---|---|---|
| Productivity impact | |||||
| Employee acceptance | |||||
| Quality improvement | |||||
| Implementation cost (higher score = more affordable) | |||||
| Workforce capability built | |||||
| Risk containment (higher score = lower risk) | |||||
| Long-term sustainability |
With your current weights, alternative A scores highest. A matrix only reflects the weights you chose. Evidence, ethics, feasibility and the people affected still belong in your judgement — you decide.