Call monitoring and feedback reporting using machine learning

1. A call monitoring device, comprising:

a network interface configured to receive phone calls; and

a processor operably coupled to the network interface, the processor configured to:

obtain at least a portion of a phone call, wherein the phone call is associated with (a) metadata that identifies keywords used during the phone call and characteristics of a voice signal associated with the phone call; and (b) a transcript for the phone call, the transcript comprising text;

input the transcript and the metadata into a machine learning model, wherein:

the machine learning model is configured to generate a call profile based on text from the transcript and the metadata for the phone call; and

the call profile comprises a first call classification for the phone call;

receive the call profile from the machine learning model in response to inputting the transcript and the metadata into the machine learning model;

identify a call log associated with the phone call, wherein the call log comprises a second call classification for the phone call;

compare the first call classification to the second call classification;

determine the first call classification does not match the second call classification;

generate a feedback report that identifies the first call classification in response to determining that the first call classification does not match the second call classification; and

output the feedback report.

2. The device of claim 1 , wherein:

the call profile further comprises one or more priority flags that are each associated with an issue type; and

the processor is further configured to:

determine the call profile comprises at least one priority flag that is set;

forward the call profile to a network device associated with a quality control unit;

receive user feedback from the network device in response to forwarding the call profile to the network device; and

populate the feedback report based on the user feedback.

3. The device of claim 1 , wherein:

the call profile further comprises one or more priority flags that are each associated with an issue type; and

the processor is further configured to:

determine the call profile comprises at least one priority flag that is set;

forward the call profile to a network device associated with a quality control unit;

receive user feedback from the network device in response to forwarding the call profile to the network device;

determine the first call classification is incorrect based on the user feedback; and

generate training data for the machine learning model based on the user feedback.

4. The device of claim 1 , wherein outputting the feedback report comprises:

identifying a group within an enterprise based on keywords used in the phone call; and

sending the call profile to a network device associated with the identified group.

5. The device of claim 1 , wherein outputting the feedback report comprises sending the feedback report to a network device.

6. The device of claim 1 , wherein the characteristics of the voice signal comprises a speech tone for a person on the phone call.

7. The device of claim 1 , wherein the characteristics of the voice signal comprise a speech rate for a person on the phone call.

8. The device of claim 1 , wherein obtaining the phone call comprises obtaining a recording of the phone call.

9. A call monitoring method, comprising:

obtaining at least a portion of a phone call, wherein the phone call is associated with (a) metadata that identifies keywords used during the phone call and characteristics of a voice signal associated with the phone call; and (b) a transcript for the phone call, the transcript comprising text;

inputting the transcript and the metadata into a machine learning model wherein:

the machine learning model is configured to generate a call profile based on text from the transcript and the metadata for the phone call; and

the call profile comprises a first call classification for the phone call;

receiving the call profile from the machine learning model in response to inputting the transcript and the metadata into the machine learning model;

identifying a call log associated with the phone call, wherein the call log comprises a second call classification for the phone call;

comparing the first call classification to the second call classification;

determining the first call classification does not match the second call classification;

generating a feedback report that identifies the first call classification in response to determining that the first call classification does not match the second call classification; and

outputting the feedback report.

10. The method of claim 9 , wherein:

the call profile further comprises one or more priority flags that are each associated with an issue type; and

further comprising:

determining the call profile comprises at least one priority flag that is set;

forwarding the call profile to a network device associated with a quality control unit;

receiving user feedback from the network device in response to forwarding the call profile to the network device; and

populating the feedback report based on the user feedback.

11. The method of claim 9 , wherein:

the call profile further comprises one or more priority flags that are each associated with an issue type; and

further comprising:

determining the call profile comprises at least one priority flag that is set;

forwarding the call profile to a network device associated with a quality control unit;

receiving user feedback from the network device in response to forwarding the call profile to the network device;

determining the first call classification is incorrect based on the user feedback; and

generating training data for the machine learning model based on the user feedback.

12. The method of claim 9 , wherein outputting the feedback report comprises:

identifying a group within an enterprise based on keywords used in the phone call; and

sending the call profile to a network device associated with the identified group.

13. The method of claim 9 , wherein outputting the feedback report comprises sending the feedback report to a network device.

14. The method of claim 9 , wherein the characteristics of the voice signal comprises a speech tone for a person on the phone call.

15. The method of claim 9 , wherein the characteristics of the voice signal comprise a speech rate for a person on the phone call.

16. The method of claim 9 , wherein obtaining the phone call comprises obtaining a recording of the phone call.

17. A computer program comprising executable instructions stored in a non-transitory computer readable medium that when executed by a processor causes the processor to:

obtain at least a portion of a phone call, wherein the phone call is associated with (a) metadata that identifies keywords used during the phone call and characteristics of a voice signal associated with the phone call; and (b) a transcript for the phone call, the transcript comprising text;

input the transcript and the metadata into a machine learning model wherein:

the machine learning model is configured to generate a call profile based on text the transcript and the metadata for the phone call; and

the call profile comprises a first call classification for the phone call;

receive the call profile from the machine learning model in response to inputting the transcript and the metadata into the machine learning model;

identify a call log associated with the phone call, wherein the call log comprises a second call classification for the phone call;

compare the first call classification to the second call classification;

determine the first call classification does not match the second call classification;

generate a feedback report that identifies the first call classification in response to determining that the first call classification does not match the second call classification; and

output the feedback report.

18. The computer program of claim 17 , wherein:

the call profile further comprises one or more priority flags that are each associated with an issue type; and

further comprising instructions that when executed by the processor causes the processor to:

determine the call profile comprises at least one priority flag that is set;

forward the call profile to a network device associated with a quality control unit;

receive user feedback from the network device in response to forwarding the call profile to the network device; and

populate the feedback report based on the user feedback.

19. The computer program of claim 17 , wherein:

the call profile further comprises one or more priority flags that are each associated with an issue type; and

further comprising instructions that when executed by the processor causes the processor to:

determine the call profile comprises at least one priority flag that is set;

forward the call profile to a network device associated with a quality control unit;

receive user feedback from the network device in response to forwarding the call profile to the network device;

determine the first call classification is incorrect based on the user feedback; and

generate training data for the machine learning model based on the user feedback.

20. The computer program of claim 17 , wherein outputting the feedback report comprises sending the feedback report to a network device.