Providing a conversational digital survey by generating digital survey questions based on digital survey responses
The present disclosure is directed toward systems, methods, and non-transitory computer readable media for generating conversational survey questions. The systems and methods analyze a received survey question response to identify characteristics of a survey response, including topics and other response features. For example, the systems can determine a sentiment associated with a given product or service that a respondent expresses within a response. Based on the determined sentiment, and further based on a set of logic rules received from a survey administrator, the systems and methods generate provide conversational follow-up questions associated with the identified product or service.
1. A method comprising:
providing, to an administrator client device, a survey creation interface comprising a plurality of manipulable survey creation blocks arranged in a nested visual survey flow mapping logical conditions defining branches of respective processes for administering a digital survey;
in response to a first client device interaction to adjust the plurality of manipulable survey creation blocks within the survey creation interface, modifying the nested visual survey flow to include a machine learning block that maps a sentiment prediction machine learning model to a branch of the logical conditions for administering the digital survey;
providing a digital survey question of the digital survey to a respondent client device;
receiving, from the respondent client device, a text-based response to the digital survey question of the digital survey;
analyzing, based on detecting that the branch for the machine learning block within the nested visual survey flow is triggered according to the logical conditions, the text-based response utilizing the sentiment prediction machine learning model of the machine learning block to identify an indicated topic and to determine a sentiment score associated with the indicated topic;
determining that the sentiment score for the indicated topic satisfies a logical condition from among the logical conditions mapped by the nested visual survey flow;
selecting, based on determining that the sentiment score for the indicated topic satisfies the logical condition, a follow-up digital survey question according to the nested visual survey flow arranged via the survey creation interface; and
providing the follow-up digital survey question to the respondent client device.
2. The method of claim 1 , wherein selecting the follow-up digital survey question comprises triggering the follow-up digital survey question according to the nested visual survey flow in response to satisfying the logical condition.
3. The method of claim 1 , wherein analyzing the text-based response to determine the sentiment score comprises utilizing the sentiment prediction machine learning model to generate the sentiment score by predicting a sentiment toward the indicated topic based on words used in the text-based response.
4. The method of claim 1 , wherein the nested visual survey flow maps the branches of respective processes for administering the digital survey by:
mapping a first possible follow-up digital survey question that is triggered in response to determining that the sentiment score is within a first range of sentiment scores; and
mapping a second possible follow-up digital survey question that is triggered in response to determining that the sentiment score is within a second range of sentiment scores.
5. The method of claim 1 , further comprising:
determining, based on analyzing the text-based response utilizing the sentiment prediction machine learning model according to the nested visual survey flow, an additional sentiment score for an additional topic within the text-based response; and
selecting an additional follow-up digital survey question to provide based on determining that the additional sentiment score satisfies an additional logical condition of the nested visual survey flow arranged via the survey creation interface.
6. The method of claim 1 , wherein determining that the sentiment score for the indicated topic satisfies the logical condition comprises determining that the sentiment score falls within a range from among a set of possible ranges of sentiment scores.
7. The method of claim 1 , further comprising:
determining a magnitude for the text-based response according to the nested visual survey flow arranged via the survey creation interface, wherein the magnitude reflects an effort taken to generate the text-based response; and
generating an overall score for the text-based response by combining the magnitude with the sentiment score according to the nested visual survey flow defining the respective processes for administering the digital survey.
8. The method of claim 7 , wherein selecting the follow-up digital survey question is further based on determining that the overall score is within a particular range of overall scores according to the nested visual survey flow.
9. A system comprising:
at least one processor; and
a non-transitory computer readable medium comprising instructions that, when executed by the at least one processor, cause the system to:
provide, to an administrator client device, a survey creation interface comprising a plurality of manipulable survey creation blocks arranged in a nested visual survey flow mapping logical conditions defining branches of respective processes for administering a digital survey;
in response to a first client device interaction to adjust the plurality of manipulable survey creation blocks within the survey creation interface, modify the nested visual survey flow to include a machine learning block that maps a sentiment prediction machine learning model to a branch of the logical conditions for administering the digital survey;
provide a digital survey question of the digital survey to a respondent client device;
receive, from the respondent client device, a text-based response to the digital survey question of the digital survey;
analyze, based on detecting that the branch for the machine learning block within the nested visual survey flow is triggered according to the logical conditions, the text-based response utilizing the sentiment prediction machine learning model of the machine learning block to identify an indicated topic and to determine a sentiment score associated with the indicated topic;
determine that the sentiment score for the indicated topic satisfies a logical condition from among the logical conditions mapped by the nested visual survey flow;
select, based on determining that the sentiment score for the indicated topic satisfies the logical condition, a follow-up digital survey question according to the nested visual survey flow arranged via the survey creation interface; and
provide the follow-up digital survey question to the respondent client device.
10. The system of claim 9 , further comprising instructions that, when executed by the at least one processor, cause the system to select the follow-up digital survey question by triggering the follow-up digital survey question according to the nested visual survey flow in response to satisfying the logical condition.
11. The system of claim 9 , further comprising instructions that, when executed by the at least one processor, cause the system to analyze the text-based response to determine the sentiment score by utilizing the sentiment prediction machine learning model to generate the sentiment score by predicting a sentiment toward the indicated topic based on words used in the text-based response.
12. The system of claim 9 , wherein the nested visual survey flow maps the branches of respective processes for administering the digital survey by:
mapping a first possible follow-up digital survey question that is triggered in response to determining that the sentiment score is within a first range of sentiment scores; and
mapping a second possible follow-up digital survey question that is triggered in response to determining that the sentiment score is within a second range of sentiment scores.
13. The system of claim 9 , further comprising instructions that, when executed by the at least one processor, cause the system to determine that the sentiment score for the indicated topic satisfies the logical condition by determining that the sentiment score falls within a range from among a set of possible ranges of sentiment scores.
14. The system of claim 9 , further comprising instructions that, when executed by the at least one processor, cause the system to:
determine a magnitude for the text-based response according to the nested visual survey flow arranged via the survey creation interface, wherein the magnitude reflects an effort taken to generate the text-based response; and
generate an overall score for the text-based response by combining the magnitude with the sentiment score according to the nested visual survey flow defining the respective processes for administering the digital survey.
15. The system of claim 14 , further comprising instructions that, when executed by the at least one processor, cause the system to select the follow-up digital survey question further based on determining that the overall score is within a particular range of overall scores according to the nested visual survey flow.
16. A non-transitory computer readable medium comprising instructions that, when executed by at least one processor, cause a computer device to:
provide, to an administrator client device, a survey creation interface comprising a plurality of manipulable survey creation blocks arranged in a nested visual survey flow mapping logical conditions defining branches of respective processes for administering a digital survey;
in response to a first client device interaction to adjust the plurality of manipulable survey creation blocks within the survey creation interface, modifying the nested visual survey flow to include a machine learning block that maps a sentiment prediction machine learning model to a branch of the logical conditions for administering the digital survey;
provide a digital survey question of the digital survey to a respondent client device;
receive, from the respondent client device, a text-based response to the digital survey question of the digital survey;
analyze, based on detecting that the branch for the machine learning block within the nested visual survey flow is triggered according to the logical conditions, the text-based response utilizing the sentiment prediction machine learning model of the machine learning block to identify an indicated topic and to determine a sentiment score associated with the indicated topic;
determine that the sentiment score for the indicated topic satisfies a logical condition from among the logical conditions mapped by the nested visual survey flow;
select, based on determining that the sentiment score for the indicated topic satisfies the logical condition, a follow-up digital survey question according to the nested visual survey flow arranged via the survey creation interface; and
provide the follow-up digital survey question to the respondent client device.
17. The non-transitory computer readable medium of claim 16 , further comprising instructions that, when executed by the at least one processor, cause the computer device to select the follow-up digital survey question by triggering the follow-up digital survey question according to the nested visual survey flow in response to satisfying the logical condition.
18. The non-transitory computer readable medium of claim 16 , further comprising instructions that, when executed by the at least one processor, cause the computer device to analyze the text-based response to determine the sentiment score by utilizing the sentiment prediction machine learning model to generate the sentiment score by predicting a sentiment toward the indicated topic based on words used in the text-based response.
19. The non-transitory computer readable medium of claim 16 , wherein the nested visual survey flow maps the branches of respective processes for administering the digital survey by:
mapping a first possible follow-up digital survey question that is triggered in response to determining that the sentiment score is within a first range of sentiment scores; and
mapping a second possible follow-up digital survey question that is triggered in response to determining that the sentiment score is within a second range of sentiment scores.
20. The non-transitory computer readable medium of claim 16 , further comprising instructions that, when executed by the at least one processor, cause the computer device to determine that the sentiment score for the indicated topic satisfies the logical condition by determining that the sentiment score falls within a range from among a set of possible ranges of sentiment scores.