At the moment, Amazon’s Rekognition service is limited to image and video analysis, but it appears to be a significant step forward as regards meaningful and intelligent metadata for unstructured content. Could save a lot of work as it uses machine learning (ML) and can process imagery in real-time.