Classifying a digital speech sample of a call to determine routing for the call

1. A computer program product for routing a call from a caller to a call agent at a call center, wherein the computer program product comprises a computer readable storage medium having program instructions embodied therewith that when executed cause operations, the operations comprising:

processing digital speech samples of call agents to categorize speech as having call agent speech attributes;

receiving information on a caller speech attribute initiating a call to the call center, wherein the caller speech attribute results from categorizing a caller digital speech sample of the caller as having the caller speech attribute;

determining a set of call agents having call agent speech attributes related to the caller speech attribute;

determining whether the caller speech attribute is categorized as being distressed;

selecting a call agent, from the set of call agents, having call agent information satisfying a criteria with respect to the caller being distressed; and

routing the call to the selected call agent to process the call.

2. The computer program product of claim 1 , wherein the operations further comprise:

storing information on the call agent speech attributes for the call agents determined by categorizing digital speech samples of the call agents in the call agent information for each of the call agents; and

storing information on the caller speech attribute, determined by processing the caller digital speech sample, in caller profile information,

wherein the selecting the call agent comprises processing information on the caller speech attribute in the caller profile information to determine a matching call agent speech attribute in the caller profile information to select the call agent.

3. The computer program product of claim 2 , wherein the caller profile information further includes information on a caller interest, and wherein the call agent information further includes information on an area of expertise, wherein the selecting the call agent further comprises:

processing information on the caller interest in the caller profile information to determine call agent profile information having an area of expertise matching the caller interest, wherein the selected call agent has a matching call agent speech attribute and a matching area of expertise to the caller speech attribute and the caller interest, respectively.

4. The computer program product of claim 1 , wherein the caller and the call agent speech attributes include at least one of regional speech dialect and speech patterns.

5. The computer program product of claim 1 , wherein the caller speech attribute indicates an emotion of the speech for the call being routed, and wherein the emotion of the speech is used to determine whether the caller speech attribute is categorized as being distressed.

6. The computer program product of claim 5 , wherein the call agent information satisfies the criteria with respect to the caller being distressed in response to at least one of:

call agent profile information indicates the call agent has expertise in handling distressed callers; and

the call agent that has worked a fewest number of hours in a current agent shift.

7. The computer program product of claim 5 , wherein the determining the emotion of the speech comprises at least one of:

using a speech analysis program to process the caller digital speech sample to determine the emotion from a sound wave encoded in the caller digital speech sample; and

converting the caller digital speech sample to text and using a tone analyzer program to determine the emotion from the text.

8. The computer program product of claim 1 , wherein the processing a digital speech sample of a call agent comprises:

converting the caller digital speech sample to text; and

processing the text, using natural language processing software, to categorize the text as a topic of a plurality of topics,

wherein the selecting the call agent comprises processing call agent information to select a call agent having indicated expertise for the topic of the text.

9. The computer program product of claim 1 , wherein the operations further comprise:

determining that there is no call agent having a call agent speech attribute that matches the caller speech attribute;

selecting an available call agent in response to determining that there is no call agent having a speech attribute matching the caller speech attribute; and

forwarding information on the caller speech attribute to the selected available call agent to use as hints to assist with the call.

10. The computer program product of claim 1 , wherein the operations further comprise:

determining that there is no call agent having a speech attribute that matches the caller speech attribute;

selecting an available call agent in response to determining that there is no call agent having a speech attribute matching the caller speech attribute;

converting, using a voice-to-text converter program, the caller digital speech sample to text;

converting, using a voice synthesizer program, the text to a digital audio file having the speech consistent with the speech attribute of the selected available call agent; and

forwarding the digital audio file to the selected available call agent to use to assist with the call.

11. A system for routing a call from a caller to a call agent at a call center, comprising:

a processor; and

a computer readable storage medium having program instructions embodied therewith that when executed by the processor cause operations, the operations comprising:

processing digital speech samples of call agents to categorize speech as having call agent speech attributes;

receiving information on a caller speech attribute initiating a call to the call center, wherein the caller speech attribute results from categorizing a caller digital speech sample of the caller as having the caller speech attribute;

determining a set of call agents having call agent speech attributes related to the caller speech attribute;

determining whether the caller speech attribute is categorized as being distressed;

selecting a call agent, from the set of call agents, having call agent information satisfying a criteria with respect to the caller being distressed; and

routing the call to the selected call agent to process the call.

12. The system of claim 11 , wherein the operations further comprise:

storing information on the call agent speech attributes for the call agents determined by categorizing digital speech samples of the call agents in the call agent information for each of the call agents; and

storing information on the caller speech attribute, determined by processing the caller digital speech sample, in caller profile information,

wherein the selecting the call agent comprises processing information on the caller speech attribute in the caller profile information to determine a matching call agent speech attribute in the caller profile information to select the call agent.

13. The system of claim 11 , wherein the caller speech attribute indicates an emotion of the speech for the call being routed, and wherein the emotion of the speech is used to determine whether the caller speech attribute is categorized as being distressed.

14. The system of claim 13 , wherein the call agent information satisfies the criteria with respect to the caller being distressed in response to at least one of:

the call agent information indicates the call agent has expertise in handling distressed callers; and

the call agent that has worked a fewest number of hours in a current agent shift.

15. The system of claim 11 , wherein the processing a digital speech sample of a call agent comprises:

converting the caller digital speech sample to text; and

processing the text, using natural language processing software, to categorize the text as a topic of a plurality of topics,

wherein the selecting the call agent comprises processing call agent information to select a call agent having indicated expertise for the topic of the text.

16. The system of claim 11 , wherein the operations further comprise:

determining that there is no call agent having a call agent speech attribute that matches the caller speech attribute;

selecting an available call agent in response to determining that there is no call agent having a speech attribute matching the caller speech attribute; and

forwarding information on the caller speech attribute to the selected available call agent to use as hints to assist with the call.

17. A computer implemented method for routing a call from a caller to a call agent at a call center, comprising:

processing digital speech samples of call agents to categorize speech as having call agent speech attributes;

receiving information on a caller speech attribute initiating a call to the call center, wherein the caller speech attribute results from categorizing a caller digital speech sample of the caller as having the caller speech attribute;

determining a set of call agents having call agent speech attributes related to the caller speech attribute;

determining whether the caller speech attribute is categorized as being distressed;

selecting a call agent, from the set of call agents, having call agent information satisfying a criteria with respect to the caller being distressed; and

routing the call to the selected call agent to process the call.

18. The method of claim 17 , further comprising:

storing information on the call agent speech attributes for the call agents determined by categorizing digital speech samples of the call agents in the call agent information for each of the call agents; and

storing information on the caller speech attribute, determined by processing the callerdigital speech sample, in caller profile information,

wherein the selecting the call agent comprises processing information on the caller speech attribute in the caller profile information to determine a matching call agent speech attribute in the caller profile information to select the call agent.

19. The method of claim 17 , wherein the caller speech attribute indicates an emotion of the speech for the call being routed, and wherein the emotion of the speech is used to determine whether the caller speech attribute is categorized as being distressed.

20. The method of claim 19 , wherein the call agent information satisfies the criteria with respect to the caller being distressed in response to at least one of:

call agent profile information indicates the call agent has expertise in handling distressed callers; and

the call agent that has worked a fewest number of hours in a current agent shift.

21. The method of claim 17 , wherein the processing a digital speech sample of a call agent comprises:

converting the caller digital speech sample to text; and

processing the text, using natural language processing software, to categorize the text as a topic of a plurality of topics,

wherein the selecting the call agent comprises processing call agent information to select a call agent having indicated expertise for the topic of the text.

22. The method of claim 17 , further comprising:

determining that there is no call agent having a call agent speech attribute that matches the caller speech attribute;

selecting an available call agent in response to determining that there is no call agent having a speech attribute matching the caller speech attribute; and

forwarding information on the caller speech attribute to the selected available call agent to use as hints to assist with the call.