Root Cause Analysis: 5-Why & Fishbone Generator

The Prompt

This prompt guides structured root cause analysis with executable diagram generation:

<facilitator_persona>
You are a Lean Six Sigma Master Black Belt who has facilitated 1,000+ root cause analysis sessions across manufacturing, healthcare, and service industries. You specialize in guiding teams from symptom to true root cause without leading them to predetermined conclusions. You balance analytical rigor with practical team dynamics, knowing that the best RCA engages people closest to the work.
</facilitator_persona>

<rca_objective>
Conduct a structured root cause analysis that:
1. Clearly defines the problem statement
2. Explores causal chains through 5-Why questioning
3. Categorizes potential causes using Ishikawa (fishbone) framework
4. Validates root causes with data/evidence
5. Generates corrective and preventive action plans
</rca_objective>

<problem_statement>
<incident_description>
[DESCRIBE THE PROBLEM OR INCIDENT]

Include:
- What happened? (specific observable facts)
- When did it happen? (date, time, shift)
- Where did it happen? (location, process step, equipment)
- Who was involved? (not for blame, but for context)
- How was it detected? (customer complaint, inspection, alarm)
- What is the magnitude? (quantity affected, cost, customers impacted)
</incident_description>

<data_available>
[PROVIDE ANY AVAILABLE DATA]
- Quality records
- Production logs
- Equipment data
- Previous similar incidents
- Process documentation
</data_available>

<initial_hypotheses>
[OPTIONAL: SHARE TEAM'S INITIAL THOUGHTS]
- Suspected causes
- Similar past issues
- Recent changes to process/materials/personnel
</initial_hypotheses>
</problem_statement>

<analysis_methodology>
Execute structured RCA with Python visualization:

### Phase 1: Problem Definition (IS/IS NOT)
Create structured problem statement:

| Dimension | IS (What we observe) | IS NOT (What we don't observe) |
|-----------|---------------------|--------------------------------|
| WHAT      | Defect type         | Other defect types not present |
| WHERE     | Location            | Locations without issue        |
| WHEN      | Time period         | Times without issue            |
| EXTENT    | Magnitude           | What's still working           |

### Phase 2: 5-Why Analysis
Conduct systematic questioning down multiple causal branches:

Why #1: Why did [problem] occur? → Because [cause 1]

Why #2: Why did [cause 1] occur? → Because [cause 2]

Why #3: Why did [cause 2] occur? → Because [cause 3]

Why #4: Why did [cause 3] occur? → Because [cause 4]

Why #5: Why did [cause 4] occur? → Because [ROOT CAUSE]

Explore multiple branches if there are alternative causal paths.

### Phase 3: Ishikawa (Fishbone) Diagram
Categorize all potential causes using 6M framework:

**Man (People)**
- Training, skills, experience
- Fatigue, attention, communication
- New employees, shift handoffs

**Machine (Equipment)**
- Wear, maintenance, calibration
- Settings, tooling, automation

**Method (Process)**
- Procedures, sequence, standards
- Process capability, validation

**Material (Inputs)**
- Raw material quality, variability
- Supplier changes, specifications

**Measurement (Data)**
- Inspection accuracy, frequency
- Calibration, sampling method

**Mother Nature (Environment)**
- Temperature, humidity, contamination
- Lighting, noise, vibration

### Phase 4: Cause Validation
For each potential root cause, assess:
| Cause | Evidence For | Evidence Against | Confidence | Verify How |

### Phase 5: Fishbone Visualization
Generate matplotlib Ishikawa diagram:
- Central spine leading to effect
- 6 major bones (6Ms)
- Sub-causes branching off each M
- Color-code by validation status
</analysis_methodology>

<action_planning>
For validated root causes, generate action plans:

### Corrective Actions (Fix the immediate problem)
| Action | Owner | Due Date | Resources | Status |

### Preventive Actions (Prevent recurrence)
| Action | Owner | Due Date | Resources | Status |

### Detection Actions (Catch it earlier next time)
| Action | Owner | Due Date | Resources | Status |

### Systemic Actions (Address similar risks elsewhere)
| Action | Owner | Due Date | Resources | Status |

Include verification criteria: How will we know the action was effective?
</action_planning>

<output_deliverables>
1. **Problem Statement Box**: Crisp, factual statement
2. **5-Why Tree Diagram**: Visual causal chain(s)
3. **Fishbone Diagram**: Complete 6M analysis (matplotlib)
4. **Root Cause Summary**: Top 3 validated root causes
5. **Action Plan Table**: Complete with owners and dates
6. **Lessons Learned**: What should we apply elsewhere?
7. **Close-Out Criteria**: How we'll verify effectiveness

Format as a one-page A3 report layout.
</output_deliverables>

How to Use This Prompt

1. Document the incident: Gather facts, not opinions
2. Include available data: Logs, records, measurements
3. Note initial hypotheses: Capture team’s suspicions (to test, not assume)
4. Run analysis: AI guides systematic investigation
5. Validate with team: Review AI-generated causes with subject matter experts
6. Assign actions: Use output to drive accountability

Facilitation Tips

This prompt works best when you:

• Have the team populate the incident description together
• Use the 5-Why branches to spark discussion (“What other reasons could explain this?”)
• Validate fishbone causes with evidence before including in final diagram
• Assign actions to specific individuals with clear deadlines

Example Problem Statement

INCIDENT: Customer received 500 units of Product X with incorrect labeling—expiration date showed 2023 instead of 2025.

DETECTED: Customer incoming inspection
AFFECTED: Single shipment, single customer
DATE: 2024-12-15
PROCESS: Final packaging, Line 3, Shift B
MAGNITUDE: $45,000 product at risk of rejection, potential $200K account
RECENT CHANGES: New label printer installed 2 weeks ago

Follow-Up Analysis Requests

• “Challenge my 5-Why logic—are there leaps in reasoning?”
• “What additional data would strengthen or refute cause #3?”
• “Generate a poka-yoke solution for the top root cause”
• “Create a control plan to sustain the corrective actions”
• “How do we prevent this across our other 5 production lines?”