You put a PDF in AWS S3 and say what to pull — the scenario reads the document and returns the data for your query.