Content generation service from video conference content for a content collaboration platform
Embodiments described herein relate to systems and methods for collaboration content creation from video conferencing content. In one or more examples, a content collaboration system may receive a notification from a video conferencing platform that a video conference session has ended. The system then receives a transcript of the video conference session. From the transcript, the system generates a transcript-processing prompt to provide to a generative output engine, which returns a generative response including a natural language string that is based on the transcript, prompt, and an output format indicated by the prompt. The system then creates a new collaboration content object, or modifies an existing object, using at least a portion of the generative response.
1 . A computer-implemented method for collaboration content generation within a content collaboration system, the method comprising:
receiving a notification that a video conference session has concluded at a video conferencing platform;
in response to receiving the notification:
providing, to the video conferencing platform using a first application program interface call, a request for a transcript of the video conference session; and
obtaining the transcript from the video conferencing platform, the transcript including natural language text generated from audio of the video conference session; and
in response to receiving the notification and the transcript:
generating a transcript-processing prompt comprising at least a portion of the natural language text of the transcript and an indication of an output format for a generative response;
providing, to a generative output engine using a second application program interface call, the transcript-processing prompt;
obtaining the generative response from the generative output engine responsive to the second application program interface call, the generative response including a natural language string according to the output format; and
causing, within a platform of the content collaboration system, generation or modification of a collaboration content object of the platform that includes at least a portion of the generative response.
2 . The computer-implemented method of claim 1 , further comprising:
in response to receiving the notification and the transcript:
processing the generative response to determine that one or more portions of the generative response are mappable to one or more references of the platform of the content collaboration system; and
replacing each portion of the one or more portions of the natural language string with a respective reference of the one or more references, wherein the one or more references comprise a person name, a team name, a page, a ticket, a board, or a card of the content collaboration system.
3 . The computer-implemented method of claim 2 , further comprising:
in response to receiving the notification:
selecting, based at least in part on a meeting type for the video conference session, a first transcript-processing prompt template from a plurality of transcript-processing prompt templates, each transcript-processing prompt template associated with one of a plurality of meeting types; and
generating the transcript-processing prompt according to the first transcript-processing prompt template.
4 . The computer-implemented method of claim 3 , further comprising:
causing generation of a graphical user interface of the content collaboration system, the graphical user interface including one or more user inputs to accept a user indication of the output format for the generative response, wherein the output format comprises one or more information presentation types, the one or more information presentation types for the transcript including one or more of a list of key points, a list of action items, a summary page, or a summary paragraph.
5 . The computer-implemented method of claim 2 , wherein processing the generative response further comprises:
determining a set of names associated with a collaboration space for the video conference session, the set of names including one or more of the person name or the team name, wherein the one or more references comprise the set of names associated with the collaboration space.
6 . The computer-implemented method of claim 2 , wherein processing the generative response further comprises:
identifying that a portion of the natural language string is a name reference that is mappable to a plurality of name references, wherein the name reference is the person name or the team name; and
selecting one name from the plurality of name references based at least in part on a context of the video conference session.
7 . The computer-implemented method of claim 6 , wherein the context of the video conference session comprises one or more of an individual participant, a team participant, a type of meeting, or a past meeting participation.
8 . The computer-implemented method of claim 1 , further comprising:
generating, wherein the transcript-processing prompt comprises a first transcript-processing prompt and the generative response comprises a first generative response, the transcript-processing prompt by splitting the natural language text of the transcript into a plurality of portions, wherein the at least a portion of the natural language text of the transcript comprises a first portion of the plurality of portions; and
in response to obtaining the first generative response:
generating a second transcript-processing prompt comprising the first generative response and a second portion of the plurality of portions;
providing, to the generative output engine using a third application program interface call, the second transcript-processing prompt;
obtaining a second generative response from the generative output engine responsive to the third application program interface call; and
combining at least the first generative response and the second generative response to generate the collaboration content object.
9 . The computer-implemented method of claim 8 , wherein the first generative response comprises a first summary of the first portion of the natural language text, and the second generative response comprises a second summary of the second portion of the natural language text.
10 . The computer-implemented method of claim 1 , further comprising:
identifying an existing content object of the content collaboration system that is referred to in the transcript; and
causing generation of a link to the existing content object within the collaboration content object that includes at least the portion of the generative response.
11 . The computer-implemented method of claim 1 , further comprising:
identifying an existing content object of the content collaboration system that is referred to in the transcript; and
inserting at least a portion of content of the existing content object into the collaboration content object that includes at least the portion of the generative response.
12 . The computer-implemented method of claim 1 , wherein:
causing modification of the collaboration content object comprises modifying one or more parameters of a task of an issue tracking platform or a documentation platform.
13 . The computer-implemented method of claim 1 , wherein:
the collaboration content object that is generated is a draft collaboration content object; and
the method further comprises:
causing generation of a graphical user interface of the content collaboration system, the graphical user interface providing a user input field for modification or confirmation of the draft collaboration content object.
14 . A content collaboration system, comprising:
a first interface configured to communicate with a video conferencing platform;
a second interface configured to communicate with a generative output engine; and
a transcript processing service coupled with the first interface and the second interface, the transcript processing service configured to:
receive, via the first interface, a notification that a video conference session has concluded at the video conferencing platform;
provide, to the video conferencing platform via the first interface and responsive to receiving the notification to the video conferencing platform, a first application program interface call that indicates a request for a transcript of the video conference session;
obtain, via the first interface and responsive to the first application program interface call, the transcript from the video conferencing platform, the transcript including natural language text generated from audio of the video conference session; and
provide, to the generative output engine via the second interface and using a second application program interface call, a transcript-processing prompt comprising at least a portion of the natural language text of the transcript;
obtain, via the second interface and responsive to the second application program interface call, a generative response including a natural language string; and
cause generation or modification of a collaboration content object for a platform of the content collaboration system, the collaboration content object including at least a portion of the generative response.
15 . The content collaboration system of claim 14 , wherein:
the transcript-processing prompt comprises at least a portion of the natural language text of the transcript and an indication of an output format for the generative response.
16 . The content collaboration system of claim 14 , further comprising:
a third interface configured to communicate with at least one client device,
wherein the transcript processing service is further configured to cause, via the third interface, generation of a graphical user interface of the content collaboration system at the at least one client device, the graphical user interface including one or more user inputs to accept a user indication of an output format for the generative response.
17 . The content collaboration system of claim 15 , wherein the output format comprises one or more information presentation types, the one or more information presentation types for the transcript including one or more of a list of key points, a list of action items, a summary page, or a summary paragraph.
18 . The content collaboration system of claim 14 , wherein:
causing modification of the collaboration content object comprises modifying one or more tasks of an issue tracking platform of the content collaboration system, or one or more pages of a documentation platform of the content collaboration system.
19 . The content collaboration system of claim 14 , wherein:
causing modification of the collaboration content object comprises updating a calendar item for the video conference session to include one or more portions of the generative response.
20 . A computer-implemented method for collaboration content generation within a content collaboration system, the method comprising:
receiving a notification that a video conference session has concluded at a video conferencing platform;
in response to receiving the notification:
providing, to the video conferencing platform using a first application program interface call, a request for a transcript of the video conference session; and
obtaining the transcript from the video conferencing platform, the transcript including natural language text for the video conference session;
causing generation of a graphical user interface of the content collaboration system, the graphical user interface including one or more user inputs to accept a user indication of an output format for a generative response for the transcript, wherein the output format comprises one or more information presentation types, including one or more of a list of key points, a list of action items, a summary page, or a summary paragraph;
in response to receiving the notification, the transcript, and the user indication of the output format:
selecting, based at least in part on a meeting type for the video conference session, a first transcript-processing prompt template from a plurality of transcript-processing prompt templates, each transcript-processing prompt template associated with one of a plurality of meeting types;
generating, according to the first transcript-processing prompt template, a transcript-processing prompt comprising at least a portion of the natural language text of the transcript and the user indication of the output format;
providing, to a generative output engine using a second application program interface call, the transcript-processing prompt;
obtaining the generative response from the generative output engine responsive to the second application program interface call, the generative response including a natural language string according to the output format and associated with a platform of the content collaboration system;
in response to receiving the generative response:
processing the generative response to determine that one or more portions of the natural language string are mappable to one or more references of the platform; and
replacing each portion of the one or more portions of the natural language string with a respective reference of the one or more references, wherein the one or more references comprise a person name, a team name, a page, a ticket, a board, or a card of the content collaboration system; and
causing generation or modification of a collaboration content object of the platform that includes the natural language string.