As part of the AI ChatBots course, I developed TelePing, an AI-driven chatbot that automates the cancellation of mobile internet subscriptions. The main goal was to explore how large language models (LLMs) could be integrated into a functional system, with the course focusing more on technical implementation than on extensive UX design.

As part of the concept phase, I created a use case flow diagram to map out the cancellation process step by step, from greeting the user to verification, cancellation, offering alternatives, and closing the interaction. This flow served as the blueprint for both the technical implementation and the conversational design.

On the technical side, I built the chatbot in Python with Streamlit as the frontend framework. I integrated OpenAI’s GPT models via LangChain, implemented prompt engineering for task-specific flows, and used ChatMemory to maintain conversational context across multiple turns. I also structured the project modularly, separating prompts, logic, and UI for clarity and maintainability. A large part of the work went into iterative testing and validation, such as refining how mobile numbers or dates were checked, to make the chatbot more reliable in handling real user input.

On the interaction design side, I kept things simple. The interface focused on guiding users step by step through verification, cancellation, and feedback. Since the course prioritized implementation, I limited the UX work to clear prompts, minimal visuals, and a straightforward flow — enough to make the chatbot usable, but not a deep dive into user research or complex design patterns.

This project gave me hands-on experience in AI integration, prompt engineering, and conversational system design. It showed how even a lightweight UI paired with solid engineering can result in a functional service prototype, and how LLMs can be steered to handle structured business tasks like subscription cancellations.