Quick answer
Check whether the AI shortlist contains suppliers that match the buying task, then verify qualification through your procurement and engineering process. For industrial parts, separate the supply role, product identity and application requirements. A supplier name in an answer is a lead for investigation; it does not establish that a part is qualified.
Details
Turn the buying task into explicit conditions
An exporter can use actual enquiries to identify the part category, application, intended delivery market and essential constraints. The following are prompt-design examples, not evidence of buyer demand or supplier performance.
| Screening task | Example question | What to check |
|---|---|---|
| Standard parts | Which suppliers offer this type of standard industrial part for this application? | Product category, model and supply role |
| Custom manufacturing | Which suppliers can evaluate a custom part against our drawing and operating conditions? | Whether custom work has been confused with selling a finished catalogue item |
| Application fit | What should we verify before shortlisting a supplier for this operating environment? | Documented conditions and unresolved requirements |
Keep category-discovery prompts free of supplier names. Run named supplier checks separately. A delivery requirement in a prompt is a request, not evidence that a supplier meets it. See the B2B question guide.
Keep a record for each candidate
Save the exact prompt, actual ChatGPT or Perplexity interface, search setting, date, full answer and visible links. Distinguish the market named in the prompt from the location or settings of the test; record anything unknown.
For each supplier, note the supply role, relevant product, public evidence and open questions. Use the same application constraints throughout the comparison. Missing information means further verification is needed, rather than proving that the supplier lacks a capability.
Separate screening from procurement approval
Record drawing review, sample validation or engineering assessment as follow-up requirements where relevant. Do not present them as completed merely because an AI answer sounds confident.
Suppliers can use this record to improve their own public explanations: identify standard products clearly, describe the confirmed scope of custom work, and state applicable limits. Publish delivery or qualification claims only when their conditions have been confirmed. Do not fill missing evidence with broad promises.
Check whether the same buying task still produces the same type of shortlist
Retain the original questions and comparable settings. Observe supplier presence, supply-role accuracy and shortlist inclusion separately. If the answer mixes unrelated types of supplier, record that mismatch before treating them as competitors. The same-conditions retest guide explains the recording method.
FAQ
Q: How should industrial parts exporters check ChatGPT and Perplexity supplier shortlists?
A: Start with a specific buying task and its requirements. Check supply role, product identity and unresolved conditions for each candidate, retaining the answer and visible sources. Treat qualification as a separate procurement step.
Q: Does an AI supplier recommendation mean a part is qualified for our application?
A: No. A recommendation identifies a candidate to investigate. Your responsible team still needs to verify the relevant product, operating conditions and any drawing, sample or engineering requirements.
Q: How do we separate custom machining suppliers from standard parts sellers in AI screening?
A: Use separate prompts for standard-part supply and custom manufacturing. Label each candidate by supply role and compare only those relevant to the same task. State delivery conditions only when confirmed.