Root cause analysis prompt

What most people type

our injection molded parts are coming out with sink marks on the thick sections. how do we figure out the root cause and fix it?

It works, but the AI has to guess your resin/grade, part description and wall thicknesses and press, mold and gate details, so the first answer usually needs another round. The version below asks for exactly what a professional would deliver.

Act as a process engineer and help me find the root cause of sink marks on thick sections of our injection molded parts, then fix them. Use a structured root cause method (a fishbone or 5 whys).

Context:
- Defect: sink marks on the thick sections of the parts
- Material: [resin/grade]
- Part and wall thickness: [part description, nominal wall and thick-section dimensions, ribs or bosses]
- Machine and tool: [press tonnage, mold type, gate type and location, cooling layout]
- Current process settings: [melt and mold temperature, injection speed, pack/hold pressure and time, cooling time, cushion]
- When it started: [always, or after a change in material, tool, machine or settings]

Include:
1. Likely causes grouped by category: material, part and tool design, process settings, machine, and environment.
2. For each cause, the data or settings to check and a quick test to confirm or rule it out, in order from fastest and cheapest to slowest.
3. Corrective actions for each confirmed cause, split into process adjustments (try first) and design or tool changes (gate size, wall thickness, ribs, cooling).
4. How to verify the fix and keep it stable, for example a short DOE or a control check.

Format: a table of cause, how to confirm and fix, followed by a short step-by-step troubleshooting order. Where my details are missing, say what you would need to know instead of guessing.

Replace the [brackets] with your details before you send it: resin/grade.

Why this version works

• Set an expert role and a structured RCA method. Gives a clear method instead of a vague 'figure out'.
• Added placeholders for resin, part, tool, settings. Sink-mark causes depend heavily on these unknown facts.
• Requested causes by category with confirming tests. Makes diagnosis testable, not just a list of guesses.
• Split fixes into process vs design changes. Lets them try cheap fixes before tooling changes.

What a professional root cause analysis prompt covers

• Structured root cause method such as fishbone or 5 whys
• Likely causes by category such as material, mold, process
• Placeholders for material and process settings
• Tests to confirm

Best AI model for this prompt

This is a hard reasoning task: multi-factor troubleshooting where wrong diagnosis wastes time, but well-known domain knowledge. You don't need the most expensive model. The cheapest model at each company that handles it well:

Pay-per-use list prices for this prompt and its expected answer (about 900 tokens). On a monthly chat plan, the same choice shows up as more messages before you hit your limit. See today's AI prices.