As of 13 August 2026, AI can only partly create a product inspection checklist.

What to actually do

1. Hand it to a person

   The route this page recommends

   A person who owns the outcome does this end to end, worth it when the failure is dear.

2. Use a tool built for this

   Second choice

3. Do it yourself

   The distant third

   A chat interface, chat-fluent skill, and roughly 30 minutes until you can act on the result.

   How to actually do it

   1. Open the current product specification, drawings, approved sample record, supplier requirements and existing inspection procedure, and collect the versions that apply to the product.
   2. Ask the person responsible for quality or engineering to identify known defects, acceptance limits, sampling rules, required test equipment and any applicable standards or customer requirements.
   3. Paste those documents or their relevant text into the prompt, replacing each bracketed slot with the product, team and inspection-stage details.
   4. Ask the chatbot to produce the checklist in a table with one check per row, including the acceptance criterion, inspection method, result field and failure action.
   5. Compare every acceptance criterion and test method in the draft with the current product documents, and remove or correct anything that has no authorised source.
   6. Give the unresolved assumptions and NEEDS HUMAN DECISION items to the quality or engineering colleague, then record their decisions in the checklist.
   7. Run the checklist against a known good product and a deliberately known defective example, record whether each check works in practice, and issue the approved version through your normal document-control process.

   Prompt

   Create a product inspection checklist for [product name] used by [inspector or team] at [inspection stage]. Use only the information I provide below and do not invent specifications, tolerances, legal requirements or test methods. Organise the checklist into clear inspection stages. For each check, include: item or feature to inspect, exact acceptance criterion, how to inspect it, equipment or reference needed, result field, pass or fail decision, and action if it fails. Include checks for identity, quantity, packaging, labelling, visible condition, dimensions or performance only where the supplied information supports them. Mark any missing requirement as NEEDS HUMAN DECISION rather than guessing. Add a short header with product version, batch or lot field, inspector, date, supplier, sampling basis and approval fields. At the end, list every assumption, ambiguity and missing document that a quality or engineering colleague must resolve before use. Source information: [paste product specification, drawings, approved sample details, known defect list, inspection procedure, sampling rules, customer requirements and applicable standards here].

   Open it prefilled in ChatGPT or Claude, or copy it into Gemini, which takes no prefill link.