Determining relevant information based on third party information and user interactions

A method comprising:

• receiving, by a background process and via one or more application programming interfaces, application data from one or more applications, the application data including features related to user activity from the one or more applications, wherein the one or more applications comprise at least one local application and at least one external application;
• providing, by the background process and using a machine learning (ML) model, a relevance score for each of one or more watch user interface (UI) elements based on the features related to the user activity;
• and
• display at least one of the one or more watch UI elements based at least in part on the relevance scores.

A computer program product comprising code stored in a non-transitory computer-readable storage medium, the code comprising:

• code to receive, by a background process and via one or more application programming interfaces, application data from one or more applications, the application data including one or more features related to user activity from the one or more applications, wherein the one or more applications comprise at least one local application and at least one external application;
• code to provide, by the background process and using a machine learning (ML) model, a respective relevance score for each of one or more user interface (UI) elements based on the one or more features;
• and
• code to display at least one of the one or more UI elements based at least in part on the respective relevance score for each of the one or more UI elements.