A factual entity is determined (404). For example, the document system 108 may identify interesting, unique, or rare terms in the content of the document. One process for identifying such terms is called Term Frequency-Inverse Document Frequency. In general, this is a statistical process that identifies a word or words that occur in a document at a much greater frequency than in a training corpus. In another example, the document system 108 may use linguistic rules and heuristics to identify a factual entity from the content. For example, in many English language texts, a subject and object of the first sentence of a paragraph may identify or be a factual entity, or used to identify a factual entity that is not explicitly stated.

Content for a knowledge panel is requested (406). For example, the document system 108 can send a request to the knowledge panel system 110 for content about the factual entity. This request may be for just the content to be displayed in a knowledge panel (e.g., text in a structured format) or may be data representing or renderable into a knowledge panel, or other appropriate data. Content for the knowledge panel is received (408). For example, the document system 108 can receive a response from the knowledge panel system 110. This response may include, for example, just content to be populated into a knowledge panel template. An example of this may be an .XML, file with text fields holding values to be filled into text fields of a template. In another example, the response may include data representing or renderable into a knowledge panel, or other appropriate data. For example, the response may contain a data object that, when processed by the document system, creates a graphical user interface object that constitutes the knowledge panel.