Skip to main content
2023 CAS Annual Meeting Main banner
Final Program


Icon Legend
This session is not in your schedule.
This session is in your schedule. Click again to remove it.
Presentation Icons
Livestreamed
Livestreamed
Not Recorded
Not Recorded
Back
Evaluation

Predictive Analytics

Level 1: No prior knowledge of the subject (0-3 Years)

C:PC:L1

C:MM:L2

C:FI:L1

C:FE:L1

C:RM:L1

C:AS:L1

S:CT:L1

S:DA:L2

S:BT:L1

S:PR:L1

S:BA:L1

S:LT:L2

T:AN:L2

T:IC:L2

T:CO:L1

T:CR:L2

T:IA:L2

T:PR:L1

CS-5 - Chat-GPT: A New Paradigm to Listen to Customers?

Tuesday, November 7, 2023
9:45 AM – 10:45 AM PST
Location: San Gabriel Ballroom (1st Floor)
With the roll-out of ChatGPT platform in late 2022, its popularity is exploding with people trying this technology to organize their thoughts, rewrite texts, or have an informed conversation around a certain topic. This session provides a case study in applying ChatGPT technology on a large scale for customer sentiment understanding. The ChatGPT approach was found to perform significantly better than the traditional lexicon-based approach in sentiment classification. In addition, it is also a great tool to extract main points from customer feedback and standardize the text in length and structure, which makes subsequent analysis (such as topic modeling, intent detection, level of urgency understanding) more effective. In this session, we also discuss how to use transformer model packages locally when passing data crossing corporate firewall is not an option.

Learning Objectives:

  • Learn how to apply the latest technologies to understand customer survey feedback on a large scale, such as using Python programming language and OpenAI API calls for ChatGPT batch-processing.
  • Understand the advantages of Chat-GPT technology over some traditional techniques for customer sentiment analysis and how this technology can be used together with other text-mining techniques to understand the main topics in customer feedback.
  • Share lessons learned in using Chat-GPT for customer sentiment analysis and information on things to avoid, limitations, etc. Discuss other alternatives such as using transformer packages locally when passing data to OpenAI is not an option.
  • Learn how to apply the latest technologies to understand customer survey feedback on a large scale, such as using Python programming language and OpenAI API calls for ChatGPT batch-processing.