Image-to-text generates natural language descriptions from images, including image captioning, OCR, and visual question answering. Multimodal models like GPT-4V and LLaVA process visual and textual information jointly. A Moroccan retail company could use image-to-text to automatically generate product descriptions from photos.