Dynamic configuration of document portions via machine learning

Techniques are described relating to dynamic document portion configuration via machine learning. An associated method includes facilitating parsing of a document to identify a plurality of document portions based upon contextual evaluation. The method further includes mining data associated with a client to determine at least one client objective associated with the document and applying at least one artificial intelligence technique of a machine learning knowledge model to the mined data in order to assign a confidence score to each of the plurality of document portions that indicates relevance of the document portion to the at least one client objective. The method further includes identifying a set of document portions among the plurality of document portions having respective confidence scores that exceed a predefined confidence score threshold, and, responsive to the client accessing the document via a user interface of a computing system, configuring the set of document portions.