one or more processors; and

computer storage memory having computer-executable instructions stored thereon which, when executed by the one or more processors, implement a method comprising:

determining a set of action item candidates from a natural language content in a transcript of content spoken during a meeting event by one or more meeting participants, wherein an individual action item candidate from the set of action item candidates includes a textual description of a task to be completed;

identifying a specific participant who spoke the natural language content through machine analysis of a meeting feature;

automatically generating a restructured action item candidate by restructuring, via a model, the textual description by removing, adding, or replacing particular words in the textual description, wherein the model uses the specific participant as an input to generate the restructured action item candidate;

detecting one or more features of the one or more meeting participants;

providing the one or more features and the restructured action item candidate as input to a machine learning model;

automatically generating a prediction estimate, by the machine-learning model processing the input, that indicates whether the restructured action item candidate is an action item or not an action item; and

based at least in part on the prediction estimate indicating that the restructured action item candidate is an action item, presenting the restructured action item candidate to one or more user devices associated with the one or more meeting participants.