Audio Content Filtering For Contact Center Agents

1 . A method, comprising:

obtaining, at a first device of a contact center agent, speech content from a second device of a contact center user during a contact center engagement between the contact center agent and the contact center user;

determining, using an artificial intelligence model accessible to the first device, that the speech content meets a threshold;

based on the speech content meeting the threshold, generating, using the artificial intelligence model, a transcription of the speech content; and

outputting, in place of the speech content and during the contact center engagement, the transcription of the speech content at the first device.

2 . The method of claim 1 , wherein determining that the speech content meets the threshold comprises:

determining that a negative emotional tone used by the contact center user within the speech content meets the threshold.

3 . The method of claim 1 , wherein determining that the speech content meets the threshold comprises:

determining that an amount of profanity used by the contact center user within the speech content meets the threshold.

4 . The method of claim 1 , wherein determining that the speech content meets the threshold comprises:

determining that a speech volume used by the contact center user within the speech content meets the threshold.

5 . The method of claim 1 , wherein outputting the transcription of the speech content at the first device comprises:

outputting, in connection with the transcription of the speech content, an indication of a negative emotional state of the contact center user.

6 . The method of claim 1 , comprising:

based on the speech content meeting the threshold, muting an audio channel of the contact center user to prevent an output of the speech content or additional content at the first device during at least some remaining amount of the contact center engagement.

7 . The method of claim 1 , wherein the artificial intelligence model is trained for sentiment analysis using contact center engagement data associated with at least one past contact center engagement for each of multiple contact center agents and the threshold is used with the contact center agent and other contact center agents.

8 . The method of claim 1 , wherein the artificial intelligence model is trained for sentiment analysis using contact center engagement data limited to the contact center agent and the threshold is specific to the contact center agent.

9 . A non-transitory computer readable medium storing instructions operable to cause one or more processors to perform operations comprising:

obtaining, at a first device of a contact center agent, speech content from a second device of a contact center user during a contact center engagement between the contact center agent and the contact center user;

determining, using an artificial intelligence model accessible to the first device, that the speech content meets a threshold;

based on the speech content meeting the threshold, generating, using the artificial intelligence model, a transcription of the speech content; and

outputting, in place of the speech content and during the contact center engagement, the transcription of the speech content at the first device.

10 . The non-transitory computer readable medium of claim 9 , wherein the threshold corresponds to one or more of a negative emotional tone, an amount of profanity, or a speech volume.

11 . The non-transitory computer readable medium of claim 9 , wherein the threshold is specific to the contact center agent.

12 . The non-transitory computer readable medium of claim 9 , wherein an indication of a negative emotional state of the contact center user is output in connection with the transcription of the speech content.

13 . The non-transitory computer readable medium of claim 9 , wherein audio from the second device is muted at the first device based on the speech content meeting the threshold.

14 . The non-transitory computer readable medium of claim 9 , wherein the speech content is obtained over a synchronous communication modality.

15 . A system, comprising:

a memory subsystem; and

processing circuitry configured to execute instructions stored in the memory subsystem to:

obtain, at a first device of a contact center agent, speech content from a second device of a contact center user during a contact center engagement between the contact center agent and the contact center user;

determine, using an artificial intelligence model accessible to the first device, that the speech content meets a threshold;

based on the speech content meeting the threshold, generate, using the artificial intelligence model, a transcription of the speech content; and

output, in place of the speech content and during the contact center engagement, the transcription of the speech content at the first device.

16 . The system of claim 15 , wherein, to determine that the speech content meets the threshold, the processing circuitry is configured to execute the instructions to:

determine that the speech content meets the threshold for a threshold period of time during the contact center engagement.

17 . The system of claim 15 , wherein, to determine that the speech content meets the threshold, the processing circuitry is configured to execute the instructions to:

determine that the speech content cumulatively meets the threshold over multiple periods of time during the contact center engagement.

18 . The system of claim 15 , wherein the processing circuitry is configured to execute the instructions to:

based on the speech content meeting the threshold, indicate a negative emotional state of the contact center user at the first device.

19 . The system of claim 15 , wherein the processing circuitry is configured to execute the instructions to:

based on the speech content meeting the threshold, mute audio of the second device.

20 . The system of claim 15 , wherein the contact center engagement is facilitated over a telephony modality or a video conferencing modality.