Questions purchasing teams ask first.
Why not upload the quotes directly to ChatGPT?
You can. QuoteCompare is intended to make the workflow repeatable: the same extraction schema, confidence gates, timestamped FX conversion, scope blockers, landed-cost rules and standardized report every time.
Does QuoteCompare choose a supplier for me?
No. It structures the evidence and highlights commercial differences. A final purchasing decision can depend on quality, supplier relationships, contractual risk and other factors outside the documents.
Does it invent missing freight, tax or customs costs?
No. Unknown monetary components remain unknown. The landed-cost engine reports a known-cost floor and flags what must be clarified.
Can it read scanned PDFs?
QuoteCompare includes OCR fallback for image-only PDF quotations. OCR accuracy can vary with scan quality, so extracted values should be checked against the source document.
Can quotes use different currencies?
Yes. Supported currencies can be normalized to a selected comparison currency using timestamped public reference rates when available.
Are my documents used to train models?
Uploaded PDFs are processed by the hosted service. When AI fallback is required, selected extracted document text may be sent to the configured OpenAI API for structured extraction. Source PDFs are removed from application storage after a successful report is generated.
Is the landed cost always final?
No. It is final only when all required components are supported by the available evidence. Otherwise the report explicitly marks the calculation incomplete.