Upload the audio as a file or as a link from AWS S3 — the scenario transcribes it, an AI writes a summary with the decisions and the tone, and sends it to Mailjet.