The scenario finds images with weak or missing captions, looks at each one, and writes a clear description — for search engines and for visitors using screen readers.