AutoBot

• 

  Landing page

• 

  Login page

• 

  Example recommendation

• 

  Recomendations And Follow Up

• 

  Architecture

This is a project for Matador.ai

AutoBot is an AI-powered car dealership assistant that makes buying a car faster and easier. It answers questions, recommends vehicles, and helps users explore various vehicles at a dealership—all through an intuitive chat interface.

What it does

AutoBot acts like a car salesperson but with fast, AI generated responses. It can:

• Answer questions about cars, features, and pricing.
• Recommend cars based on your budget and preferences.
• Help you browse available cars in the dealership inventory.

It’s like having your own car expert, 24/7.

How we built it

• Frontend: Built with Next.js for a fast and modern UI.
• Backend: Flask handles the logic, and MongoDB stores inventory and user data.
• AI: We used TensorFlow, scikit-learn, NLP, and LLMs like ChatGPT for recommendations and conversations.
• Integrated APIs to connect everything seamlessly.

Challenges we ran into

• Not knowing where to start and how to approach the problem at hand.
• Training AI models to understand car-specific questions.
• Making sure the system runs fast and handles data smoothly.
• Debugging issues between the frontend, backend, and database.

Accomplishments that we're proud of

• Building a functional AI assistant in a short time.
• Getting the AI to handle real dealership-specific queries well.
• Making the frontend sleek and easy to use.

What we learned

• How to integrate a bunch of different technologies into one system.
• Challenges of training AI for a niche use case.
• The importance of designing a good user experience.

What's next for AutoBot

• Add voice interaction for a hands-free experience
• Build a mobile app for on-the-go car browsing.
• Keep improving the AI based on user feedback.

Built With

• llm
• natural-language-processing
• numpy
• pandas
• python
• react
• scikit-learn
• spacy