Create a fishbone (Ishikawa) diagram of possible causes
Sorts every possible cause of a problem into a fishbone diagram, ranks the likely ones, and tells you how to check each.
The prompt
Help me build a fishbone (Ishikawa) diagram for a problem.
Problem (the head of the fish): [WHAT IS GOING WRONG, measurable if possible - e.g. "30% of new hires leave within 90 days"]
Where it happens: [team, process, office, product, customer type]
Data, notes, or complaints I have: [PASTE]
Causes the team already suspects: [LIST]
Steps:
1. Choose the cause categories that fit my problem. Start from People, Process/Methods, Tools/Technology, Materials/Inputs, Measurement, and Environment (or the 4 Ps for service work: Policies, Procedures, People, Place), and rename or swap categories so they fit.
2. Under each category, list 3-6 specific possible causes, with sub-causes where useful. Include causes the team hasn't mentioned.
3. Mark each cause: supported by my data, suggested by someone, or speculative.
4. Pick the 3-5 most likely causes and explain why.
Return:
1. The fishbone as a text diagram or indented list (Problem > Category > Cause > Sub-cause).
2. A Mermaid flowchart (graph LR, with the problem on the right and one branch per category) I can paste into a tool that renders Mermaid, then copy into slides.
3. A table for the top causes: Cause | Category | Evidence so far | How to verify (data to pull or person to ask) | Effort to verify.
4. What to do next: which causes to verify first, and a quick test for each.
Rules: list causes, not solutions. Don't present any cause as confirmed unless my data shows it. Keep wording neutral and free of blame toward individuals.
Highlighted [brackets] are blanks to replace with your own details.
Fill in the blanks and use this promptMore Thinking & Decisions prompts
What to fill in
- team, process, office, product, customer type
- PASTE
- LIST
How this prompt is built
This prompt follows the RTF structure (Role, Task, Format). RTF tells the model who to be, what to do and what shape the answer should take. It is the quickest structure that still gives noticeably better results than a bare request.
Related prompts
- Steelman both sides of an argumentThe strongest honest case for each side of a debate, and where the real disagreement lies.
- Brainstorm fresh ideas on any problemGenerates a wide range of ideas — including unconventional ones — then helps you pick the best.
- Find the gaps and blind spots in my report before I share itReviews your report or analysis for missing information, unanswered questions, untested assumptions, and what a skeptical reader will ask.
- Turn research findings into prioritized business recommendationsConverts research or analysis into specific, prioritized recommendations with owners, effort, impact, and how to measure success.