| CPC G10L 17/06 (2013.01) [G10L 15/26 (2013.01); H04L 65/1069 (2013.01)] | 20 Claims |
|
1. A method performed by one or more computing systems for augmenting session transcripts, the method comprising:
receiving session data comprising a set of audiovisual recording data;
receiving session metadata comprising: data derived from an environment in which the set of audiovisual recording data was captured, or data associated with the set of audiovisual recording data, or both;
parsing the session data and the session metadata to identify a set of speakers;
identifying a set of channels, wherein each channel in the set of channels corresponds to a speaker in the set of speakers;
applying a machine learning speech model to generate a rough transcript corresponding to the set of audiovisual recording data,
wherein each segment of the rough transcript corresponds to a speaker from the set of speakers and a channel from the set of channels;
for each channel in the set of channels:
identifying a set of segments of the rough transcript associated with the channel;
mapping the channel to a particular verification agent within a set of verification agents using one or more competencies of the particular verification agent,
wherein the particular verification agent mapped to each channel in the set of channels is different, and
wherein the particular verification agent is configured to edit the set of segments of the rough transcript associated with the mapped channel;
transmitting the identified set of segments to the particular verification agent; and
receiving an edited version of the identified set of segments from the particular verification agent; and
generating an augmented session transcript by merging the edited versions of the set of segments received for each channel in the set of channels.
|