Automate customer support handling E-Commerce enquiry using ChatGPT
The primary objective of this thesis is to improve customer support on e-commerce platform by proposing an innovative solution that integrates advanced technologies and methodologies. The motivation stems from the need to enhance the efficiency of the customer reply system, reduce the workload on...
Saved in:
| Main Author: | |
|---|---|
| Format: | Final Year Project / Dissertation / Thesis |
| Published: |
2024
|
| Subjects: | |
| Online Access: | http://eprints.utar.edu.my/6627/1/20ACB02426_FYP2.pdf http://eprints.utar.edu.my/6627/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | The primary objective of this thesis is to improve customer support on e-commerce platform
by proposing an innovative solution that integrates advanced technologies and methodologies.
The motivation stems from the need to enhance the efficiency of the customer reply system,
reduce the workload on the customer support team, and increase company sales. One of the
long-term goal in the field of AI is to build computer systems that can have human-like
conversations with users. With recent advances in AI technologies, we are now one step
closer to achieving this goal. This proposal makes significant contributions to reduce
manpower dependency and increase overall business competency in customer support on ecommerce
platforms. The development of an automated context handling mechanism ensures
precise and efficient customer support by reducing the need for human manpower. The
automated summarization feature streamlines human agents' tasks by summarizing the entire
conversations. By this it can help in saving time and increasing overall competency. The use
of ChatGPT enhances the competency of business interactions by providing contextually
relevant and precise responses. We are going to integrate this mechanism with e-commerce
platform aligns with evolving customer communication preferences and enhances the
business's competency in answering to customer inquiries on this e-commerce platform.
Additionally, order processing functionality will be integrated within the chat interface to
provide convenience to the customers so that they can easily make order using a shorter time.
The project scope is the development of a comprehensive mechanism for context handling, an
inconspicuous human takeover process, and the summarization of entire conversations
between customers and automated customer support before handover to human agents. The
function of context handling ensures that automated responses remain relevant to the business,
while summarization significantly can help in reducing the workload on human agents during
handover sessions. Integration with ChatGPT allows for accurate responses, and integrate
with the Instagram platform enables efficient responses to customer questions. The project
also includes the implementation of functionality for customers to place orders through the
chat interface. The methodology involves the design of a distributed system architecture for
scalability and efficient task distribution. Machine Learning-based Named Entity Recognition
(NER) is employed to identify and extract specific entities, while contextual analysis
algorithms determine message relevance for summarization. Reinforcement learning
techniques will adapt the summarization model based on human agent feedback, and
Bachelor of Information Systems (Honours) Information Systems Engineering
Faculty of Information and Communication Technology (Kampar Campus), UTAR
7
feedback analysis identifies areas for continuous improvement. This comprehensive approach
aims to transform customer support on Instagram, offering a seamless and efficient
experience that reduces workload, accelerates sales, and enhances overall business
competency. Last but not least, the core objective is to propose a new mechanism that
automatically summarizes conversations between automated customer support and customers,
facilitating effective handover to human agents. The proposed solution focuses on creating a
natural and human-like conversation flow, aligning with customer preferences, and
minimizing response times for quicker response from the customer support.
|
|---|