Training a machine learning algorithm to create survey questions
In some examples, a server may determine that a case, created to address an issue of a computing device, is closed and perform an analysis of steps in a process used to close the case. The analysis may determine a length of time of each step and determine that a time to close the case or complete a particular step was at least a predetermined amount faster than average. The server may use machine learning to create a survey question to determine a technique used to close the case or complete the particular step faster than average and to determine one or more incentives to provide a technician that closed the case. An answer from the technician to the survey question may include the technique used to close the case or complete the particular step faster than average. The technique may be shared with other technicians.
1. A method comprising:
determining, by a server associated with a technical support group, that a case created to address an issue associated with a computing device is closed, wherein determining that the created case is closed includes establishing and recording a communications session related to the case, wherein establishing the communications session includes connecting to a client device corresponding to the case, where the client device connection is established via the communications session or another parallel computing session;
performing, by the server, an analysis of a process used to close the case, the analysis comprising determining a length of time corresponding to individual steps in the process used to close the case based on telemetry information related to the recorded communications session;
determining, using a machine learning algorithm of one or more Support Vector Machines (SVMs) executed by the server and based at least in part on the analysis, that the case was closed faster than an average case time to close similar cases;
determining, using the machine learning algorithm and based at least in part on the analysis, that a particular step in the process used to close the case was performed faster than an average step time associated with the particular step;
generating, using the machine learning algorithm, a second question to include in the custom survey to determine a second technique used to perform the step faster than the average step time;
after providing the custom survey to the technician, receiving a second answer corresponding to the second question, the second answer including the second technique used to perform the step faster than the average step time;
providing the second technique used to perform the step faster than the average step time to the one or more additional technicians;
generating, using the machine learning algorithm, a custom survey including a question to determine a technique used to close the case faster than the average case time;
determining, using the machine learning algorithm, one or more incentives to provide to a technique to respond to the custom survey;
providing, by the server, the custom survey to a technician that closed the case faster than the average case time;
identifying, by the server, the one or more incentives to the technician;
receiving, by the server, an answer corresponding to question, the answer including the technique used to close the case faster than the average case time; and
providing, by the server, the technique used to close the case faster than the average case time to one or more additional technicians.
2. The method of claim 1 , wherein the analysis of the process used to close the case further comprises:
determining one or more actions performed in the individual steps in the process used to close the case.
3. The method of claim 1 , wherein the one or more incentives comprise at least one of:
a cash award;
points redeemable for goods or services; or
an added responsibility of the technician, wherein a job performance of the technician is based at least in part on an evaluation of the added responsibility.
4. The method of claim 1 , wherein the analysis of the process used to close the case further comprises performing skills-issue correlation, the skills-issue correlation comprising:
determining an issue associated with the case;
determining one or more skills used to close the case; and
determining a correlation between the one or more skills and skills used by the one or more additional technicians to address similar issues.
5. The method of claim 1 , wherein the analysis of the process used to close the case further comprises performing issue-solution correlation, the issue-solution correlation comprising:
determining an issue associated with the case;
determining a solution used to address the issue; and
determining an appropriateness of the solution.
6. The method of claim 1 , wherein determining, using the machine learning algorithm executed by the server and based at least in part on the analysis, that the case was closed faster than the average case time to close similar cases comprises:
determining a case time comprising a difference between a first point in time that the case was created to a second point in time that the case was closed;
determining the average case time to close the similar cases; and
determining that the case time is at least a predetermined percentage less than the average case time.
7. A server comprising:
one or more processors; and
one or more non-transitory computer-readable media storing instructions executable by the one or more processors to perform operations comprising:
determining that a case created to address an issue associated with a computing device is closed, wherein determining that the created case is closed includes establishing and recording a communications session related to the case, wherein establishing the communications session includes connecting to a client device corresponding to the case, where the client device connection is established via the communications session or another parallel computing session;
performing an analysis of a process used to close the case, the analysis comprising determining a length of time corresponding to individual steps in the process used to close the case based on telemetry information related to the recorded communications session;
determining, via at least one Support Vector Machine (SVM) and based at least in part on the analysis, that the case was closed faster than an average case time to close similar cases;
determining, using the machine learning algorithm and based at least in part on the analysis, that a particular step in the process used to close the case was performed faster than an average step time associated with the particular step;
generating, using the machine learning algorithm, a second question to include in the custom survey to determine a second technique used to perform the step faster than the average step time;
after providing the custom survey to the technician, receiving a second answer corresponding to the second question, the second answer including the second technique used to perform the step faster than the average step time;
providing the second technique used to perform the step faster than the average step time to the one or more additional technicians;
generating a custom survey including a question to determine a technique used to close the case faster than the average case time;
determining, using the machine learning algorithm, one or more incentives to provide to a technique to respond to the custom survey;
providing the custom survey to a technician that closed the case faster than the average case time;
identifying the one or more incentives to the technician;
receiving an answer corresponding to question, the answer including the technique used to close the case faster than the average case time; and
providing the technique used to close the case faster than the average case time to one or more additional technicians.
8. The server of claim 7 , wherein the analysis of the process used to close the case further comprises:
determining one or more actions performed in the individual steps in the process used to close the case.
9. The server of claim 7 , wherein the one or more incentives comprise at least one of:
a cash award;
points redeemable for goods or services; or
an added responsibility of the technician, wherein a job performance of the technician is based at least in part on an evaluation of the added responsibility.
10. The server of claim 7 , wherein determining, using the machine learning algorithm executed by the server and based at least in part on the analysis, that the case was closed faster than the average case time to close similar cases comprises:
determining a case time comprising a difference between a first point in time that the case was created to a second point in time that the case was closed;
determining the average case time to close the similar cases; and
determining that the case time is at least a predetermined percentage less than the average case time.
11. One or more non-transitory computer-readable media storing instructions executable by one or more processors to perform operations comprising:
determining that a case created to address an issue associated with a computing device is closed, wherein determining that the created case is closed includes establishing and recording a communications session related to the case, wherein establishing the communications session includes connecting to a client device corresponding to the case, where the client device connection is established via the communications session or another parallel computing session;
performing an analysis of a process used to close the case, the analysis comprising determining a length of time corresponding to individual steps in the process used to close the case based on telemetry information related to the recorded communications session;
determining, via at least one Support Vector Machine (SVM) and based at least in part on the analysis, that the case was closed faster than an average case time to close similar cases;
determining, using the machine learning algorithm and based at least in part on the analysis, that a particular step in the process used to close the case was performed faster than an average step time associated with the particular step;
generating, using the machine learning algorithm, a second question to include in the custom survey to determine a second technique used to perform the step faster than the average step time;
after providing the custom survey to the technician, receiving a second answer corresponding to the second question, the second answer including the second technique used to perform the step faster than the average step time;
providing the second technique used to perform the step faster than the average step time to the one or more additional technicians;
generating a custom survey including a question to determine a technique used to close the case faster than the average case time;
determining, using the machine learning algorithm, one or more incentives to provide to a technique to respond to the custom survey;
providing the custom survey to a technician that closed the case faster than the average case time;
identifying the one or more incentives to the technician;
receiving an answer corresponding to question, the answer including the technique used to close the case faster than the average case time; and
providing the technique used to close the case faster than the average case time to one or more additional technicians.
12. The one or more non-transitory computer-readable media of claim 11 , wherein the analysis of the process used to close the case further comprises:
determining one or more actions performed in the individual steps in the process used to close the case.
13. The one or more non-transitory computer-readable media of claim 11 , further comprising:
determining, using the machine learning algorithm and based at least in part on the analysis, that a particular step in the process used to close the case was performed faster than an average step time associated with the particular step;
generating, using the machine learning algorithm, a second question to include in the custom survey to determine a second technique used to perform the step faster than the average step time;
after providing the custom survey to the technician, receiving a second answer corresponding to the second question, the second answer including the second technique used to perform the step faster than the average step time; and
providing the second technique used to perform the step faster than the average step time to the one or more additional technicians.
14. The one or more non-transitory computer-readable media of claim 11 , wherein the one or more incentives comprise at least one of:
a cash award;
points redeemable for goods or services; or
an added responsibility of the technician, wherein a job performance of the technician is based at least in part on an evaluation of the added responsibility.
15. The one or more non-transitory computer-readable media of claim 11 , wherein the analysis of the process used to close the case further comprises performing skills-issue correlation, the skills-issue correlation comprising:
determining an issue associated with the case;
determining one or more skills used to close the case; and
determining a correlation between the one or more skills and skills used by the one or more additional technicians to address similar issues.
16. The one or more non-transitory computer-readable media of claim 11 , wherein the analysis of the process used to close the case further comprises performing issue-solution correlation, the issue-solution correlation comprising:
determining an issue associated with the case;
determining a solution used to address the issue; and
determining an appropriateness of the solution.
17. The one or more non-transitory computer-readable media of claim 11 , wherein determining, using the machine learning algorithm executed by the server and based at least in part on the analysis, that the case was closed faster than the average case time to close similar cases comprises:
determining a case time comprising a difference between a first point in time that the case was created to a second point in time that the case was closed;
determining the average case time to close the similar cases; and
determining that the case time is at least a predetermined percentage less than the average case time.