Methods and apparatus for determining non-textual reply content for inclusion in a reply to an electronic communication

A method implemented by one or more processors, the method comprising:

• identifying an electronic communication sent to a user;
• determining a likelihood that the electronic communication will be replied to with non-textual reply content, wherein determining the likelihood comprises:
• providing, as input to a trained machine learning system, at least one message feature of the electronic communication;
• and
• receiving, as output from the trained machine learning system that is responsive to the input, the likelihood that the electronic communication will be replied to with non-textual reply content;
• in response to the likelihood satisfying a threshold:
• causing a selectable graphical element, corresponding to at least one electronic document that is relevant to the electronic communication, to be presented at a client device of the user while the electronic communication is also being presented at the client device;
• receiving a selection of the selectable graphical element, the selection being by the user via a user interface input device of the client device;
• and
• in response to receiving the selection, performing one or both of:
• automatically attaching the at least one selected electronic document to a reply electronic communication that is a reply to the electronic communication, and
• automatically incorporating a link to the at least one selected electronic document into the reply electronic communication.

A system, comprising:

• an electronic communication stored in one or more non-transitory computer readable media, the electronic communication sent to a user;
• at least one processor;
• and
• memory coupled to the processor, wherein the memory stores instructions to be executed by the processor to perform steps comprising:
• determining a likelihood that the electronic communication will be replied to with non-textual reply content, wherein determining the likelihood comprises:
• providing, as input to a trained machine learning system, at least one message feature of the electronic communication;
• and
• receiving, as output from the trained machine learning system that is responsive to the input, the likelihood that the electronic communication will be replied to with non-textual reply content;
• in response to the likelihood satisfying a threshold:
• causing a selectable graphical element, corresponding to at least one electronic document that is relevant to the electronic communication, to be presented at a client device of the user while the electronic communication is also being presented at the client device;
• receiving a selection of the selectable graphical element, the selection being by the user via a user interface input device of the client device;
• and
• in response to receiving the selection:
• automatically attaching the at least one selected electronic document to a reply electronic communication that is a reply to the electronic communication, or
• automatically incorporating a link to the at least one selected electronic document into the reply electronic communication.