Annotation adds meaningful labels to data so that they can be used as a means of learning by various systems. It is pivotal that the data is structured in the right way to make it useable for machine learning. There are many different annotation task types, depending on project needs. Linguistic annotation tasks include morphosyntactic annotation, part-of-speech tagging, named entity annotation, and many more.
If natural language processing (NLP) is based on supervised learning, annotated/labeled data is of crucial importance. Simple examples would include:
A virtual assistant holding a conversation by tracking anaphoric usages (e.g., pronouns referring to something else in the text).
A data extraction system scanning text to retrieve the most important information for the project.
A text summarization tool to crop out unimportant parts of a text to gather important information in a more concise way.