Compliance detection using natural language processing

A method, comprising:

• transcribing a conversation to a transcript, the conversation being representative of a sound recording that occurs between at least an agent and a client;
• obtaining an action log that corresponds to the conversation, the action log identifying times in the conversation that relate to actions in a workflow;
• identifying a section of the transcript based at least in part on the action log, wherein the section of the transcript is associated with one or more topics of discussion corresponding to the workflow and based at least in part on one or more actions taken during the section of the transcript, wherein the one or more actions represent a click event, a mouse event, a keyboard event, or other computer captured event;
• obtaining a compliance rule for the section of the transcript, wherein the compliance rule is based at least in part on the one or more topics of discussion associated with the section of the transcript;
• determining, using a natural language processor, that the compliance rule has been violated for the section of the transcript;
• and
• storing a compliance violation comprising the compliance rule and the section of the transcript.

A system, comprising:

• a computing device comprising a processor and a memory;
• and
• machine-readable instructions stored in the memory that, when executed by the processor, cause the computing device to at least:
• transcribe a sound recording to a transcript, the sound recording being representative of a telephonic call occurring between at least an agent and a client;
• determine, based at least in part on an analysis by a natural language processor of the transcript, that the telephonic call included a discussion related to a regulated subject;
• obtain an action log that corresponds to the telephonic call, the action log identifying times in the telephonic call that relate to one or more actions, wherein the one or more actions represent one or more click events, one or more mouse events, one or more keyboard events, one or more computer captured events, or a combination of any thereof;
• obtain a compliance rule based at least in part on the regulated subject;
• determine, using the natural language processor of the transcript, that the compliance rule has been violated;
• and
• store at least the compliance rule that has been violated and a portion of the transcript that is representative of the compliance rule that has been violated, wherein the portion of the transcript is based at least in part on one or more actions taken during the portion of the transcript, wherein the one or more actions represent the one or more click events, the one or more mouse events, the one or more keyboard events, the one or more computer captured events, or a combination of any thereof.