Managed ML services
If you are dealing with a well-defined general-purpose ML problem in the domain of text, image, video, language, or documents, then the chances are high that Azure already provides a managed ML service for this problem.
Managed ML services are very easy to use, quick to embed into an application, and usually don't require any operational overhead. This makes them perfect for creating AI-based applications or features without the need for collecting training data, training models, and operating model deployments in production. Most importantly, managed ML services don't require any ML expertise to build ML-based applications.
Some examples of well-defined ML problems are image classification, image tagging, object detection, face detection, handwriting recognition, speech-to-text and text-to-speech conversion, speaker recognition, translation, spell-checking, keywords and entity extraction, sentiment analysis, adult content filtering, and document...