Vinny the Verizon Gator

Meet Vinny, the Verizon Gator—your 24/7 mobile plan advisor! Using generative AI, Vinny helps you navigate Verizon's plan options, answering questions and recommending the best fit for you.

Comment 1

Inspiration

The inspiration for this project came from the realization that selecting a Verizon plan can be overwhelming for users. With so many plans and options available, it's difficult to find the one that best suits individual needs. We wanted to make this process as smooth as possible.

What it does

Vinny the Verizon Gator is a chatbot designed to educate users about the various Verizon plans available and assist them in finding the most suitable one. The chatbot answers queries and makes personalized recommendations based on a sample database sourced from Verizon's FAQ page.

How we built it

We built Vinny using Python for the backend, with Flask serving as the web framework. The frontend was designed using HTML, CSS, and JavaScript. The AI capabilities are powered by a model trained through the Claude API and supplemented with short learning techniques.

Challenges we ran into

Some of the challenges included sourcing accurate and comprehensive data on Verizon's plans. We also had to ensure that the chatbot could understand and respond to a wide variety of user queries, which required fine-tuning the AI model.

Accomplishments that we're proud of

We're proud of successfully integrating the AI model into the chatbot, ensuring a smooth and user-friendly experience. The chatbot not only answers queries but also provides personalized recommendations, making it a one-stop solution for users looking for the right Verizon plan.

What we learned

Throughout this project, we learned the intricacies of chatbot development, from data sourcing to AI model training and UI design. We also gained valuable insights into user behavior and preferences, which will help us in future iterations.

What's next for Vinny the Verizon Gator

The next step for Vinny includes expanding the dataset to cover even more plans and options, as well as implementing additional features like voice recognition and multi-language support.

Built With

+ 1 more
Share this project:

Updates

posted an update —

Here are all the ways we used GenAI:

  1. The chatbot itself is an AI assistant trained on Verizon data publicly found in the FAQ. The chatbot uses an open source LLM from Google named PaLM.
  2. We used Claude, another separate LLM to generate more data as the Verizon FAQ was not sufficient, augmenting our dataset for training the PaLM model to be a better Verizon customer service rep.
  3. We used DALL.E 3 to generate the logo, and then runwayML to animate the chatbot when the bot responds.
  4. Most of us aren't computer scientists, and so chatGPT helped with integrating the frontend (HTML with JS) and backend (Flask in Python).

Log in or sign up for Devpost to join the conversation.