The scenario runs a Microsoft Teams assistant that understands text, voice notes, photos and PDFs, answers to the point, and generates images on request.