Generating test scenarios by detecting failure patterns and themes in customer experiences

Patent metadata

Field	Value
Publication number	US-11841892-B2
Application number	US-202117198836-A
Country	US
Kind code	B2
Filing date	Mar 11, 2021
Priority date	Mar 11, 2021
Publication date	Dec 12, 2023
Grant date	Dec 12, 2023

Abstract

The described technology is generally directed towards processing various customer input data to extract frequently recurring customer experience themes, including positive and negative sentiment regarding customer experiences. Natural language processing, image processing, speech recognition and/or computer vision techniques can be used on customer-related data to determine themes, tests and scenarios, as well as discover insights that can be used to improve customer experiences. The technology can be used to recreate a customer engagement, journey and overall experience by designing test scenarios around failure themes.

First claim

What is claimed is:

1. A system, comprising:
a processor, and a memory that stores executable instructions that, when executed by the processor, facilitate performance of operations, the operations comprising:
processing text data comprising respective customer verbatim input information into respective modified verbatim datasets representative of the respective customer verbatim input information;
processing the respective modified verbatim datasets into a group of topics;
determining, for the respective modified verbatim datasets, respective sentiment indicators;
adding the respective sentiment indicators to the respective modified verbatim datasets;
for each respective modified verbatim dataset, selecting a respective topic of the group of topics based on respective text in the respective modified verbatim dataset, and adding the respective topic to the respective modified verbatim dataset, the selecting and the adding for each respective modified verbatim dataset resulting in respective topics being added to the respective modified verbatim datasets;
extracting theme data from the respective modified verbatim datasets;
selecting a simulated customer persona; and
generating a test to evaluate a simulated customer experience based on the theme data and the simulated customer persona.

2. The system of claim 1 , wherein the text data comprises verbatim text data, and wherein the processing of the text data into the respective modified verbatim datasets comprises at least one of: correcting spelling errors in the verbatim text data, removing stopwords from the verbatim text data, or removing personal information from the verbatim text data.

3. The system of claim 1 , wherein the operations further comprise adding respective subtopics to the respective modified verbatim datasets based on the respective text in the respective modified verbatim datasets.

4. The system of claim 1 , wherein the theme data extracted from the respective modified verbatim datasets comprises a first part of the theme data, wherein the operations further comprise applying image processing to image data to obtain image recognition text data, and wherein the theme data comprises a second part of the theme data based on the image recognition text data.

5. The system of claim 4 , wherein the image data corresponds to a photograph.

6. The system of claim 4 , wherein the image data corresponds to a frame of video data, and wherein the theme data comprises a third part of the theme data based on recognized text-to-speech data recognized from speech data associated with the image data.

7. The system of claim 1 , wherein the processing of the respective modified verbatim datasets into the group of topics comprises topic modeling the modified verbatim datasets via an Lda2vec model.

8. The system of claim 1 , wherein the determining of the respective sentiment indicators comprises determining, for each modified verbatim dataset of the respective modified verbatim datasets, a positive sentiment indicator or negative sentiment indicator, and associating the modified verbatim dataset with the positive sentiment indicator or negative sentiment indicator.

9. The system of claim 8 , wherein the extracting of the theme data comprises grouping, based on the respective topics, the respective modified verbatim datasets into modified verbatim dataset topic groups, and obtaining a sentiment score for each topic group of the modified verbatim dataset topic groups, comprising counting associated modified verbatim datasets in the topic group that are associated with negative sentiment indicators.

10. The system of claim 9 , wherein the operations further comprise evaluating the sentiment score for each topic group relative to a threshold value to select the theme data to use for the generating of the test.

11. A method, comprising:
obtaining, by a system comprising a processor, customer input data related to customer interaction experiences;
topic modeling, by the system, the customer input data into a group of topics via natural language processing;
processing, by the system, the customer input data into datasets, in which respective datasets of the datasets comprise respective text data corresponding to respective customer input data, respective sentiment data determined from the respective text data, and respective topic data representative of at least one respective topic of the group of topics to which the respective text data threshold correlates; extracting, by the system, theme data from the datasets;
selecting, by the system, a simulated customer persona; and
generating, by the system, a test that evaluates a simulated customer experience based on the theme data and the simulated customer persona.

12. The method of claim 11 , wherein the obtaining of the customer input data related to the customer interaction experiences further comprises receiving image data, and classifying the image data.

13. The method of claim 11 , further comprising, determining, by the system based on the respective topic data and the respective text data of the respective datasets, respective subtopic data representative of at least one respective subtopic of the at least one respective topic, and associating the respective subtopic data with the respective datasets.

14. The method of claim 11 , where the extracting of the theme data from the datasets comprises processing the respective sentiment data, the respective topic data and respective subtopic data of the respective datasets to determine a representative value representing a group of datasets determined to have a common topic and a common subtopic, and determined to have negative sentiment data.

15. The method of claim 14 , further comprising triggering, by the system, the generating of the test based on the representative value being determined to have reached a threshold value.

16. The method of claim 11 , wherein the obtaining of the customer input data comprises processing customer verbatim text data into the customer input data, comprising correcting spelling errors in the verbatim text data, removing stopwords from the verbatim text data, or removing personal information from the verbatim text data.

17. The method of claim 11 , wherein the topic modeling of the customer input data via the natural language processing comprises inputting the customer input data into an Lda2vec model.

18. A non-transitory machine-readable storage medium, comprising executable instructions that, when executed by a processor, facilitate performance of operations, the operations comprising:
processing raw customer text commentary into clean text data and sentiment data, the clean text data and sentiment data comprising first text data;
modeling the first text data via natural language processing into a group of topics; associating the first text data with at least one topic of the group of topics;
processing customer image data related to customer commentary into second text data, the second text data comprising text information recognized from the customer image data;
processing customer speech data related to customer commentary into third text data;
extracting theme data representative of failure patterns from the first text data, the second text data and the third text data;
selecting a simulated customer persona; and
generating, based on the theme data and the simulated customer persona, test instructions, representative of a test, that, when executed, simulate a customer experience.

19. The non-transitory machine-readable storage medium of clai

Assignees

• Dell Products Lp

Inventors

• Mathur Prateek
• Arora Anish
• Kolodzey Mallory Anne
• Singh Shalu
• Sawhney Amit
• Bikumala Sathish Kumar
• Kaura Gautam K

Classifications

• G06F16/353 Primary

  into predefined classes · CPC title

• G06F40/166

  Editing, e.g. inserting or deleting · CPC title

• G06F40/232

  Orthographic correction, e.g. spell checking or vowelisation · CPC title

• G06F40/30

  Semantic analysis · CPC title

• G06N20/00

  Machine learning · CPC title