GIF
Demo of our chatbot waiter

Inspiration

As a team, we were inspired to help grow one of our member's father's restaurant business. We recognized that one of the biggest barriers for people exploring ethnic cuisines is often a lack of familiarity with the dishes and ingredients. This can lead to uncertainty when ordering or even deter people from trying new foods altogether. To bridge this gap and make the restaurant's delicious Indian cuisine more accessible, we came up with the idea for Naan Negotiator - an AI-powered chatbot that acts as a virtual waiter. Naan Negotiator allows customers to ask questions about the menu, get recommendations, and even place orders from the comfort of their own homes. Our goal was to recreate the experience of having a knowledgeable and friendly waiter guide you through the menu, but with the added convenience of a web app.

What We Learned

Throughout the development process, our team gained valuable skills and knowledge across multiple areas: Chatbot Development: We learned how to leverage the power of OpenAI's language models to create an intelligent, conversational chatbot capable of understanding and responding to user queries about Indian food. Web Development: Building Naan Negotiator gave us hands-on experience with Next.js, a popular React framework for creating server-rendered web applications. We learned how to structure and optimize our code for performance and scalability. Web Scraping: To train our AI assistant, we needed to gather data from the restaurant's website. We learned techniques for effectively scraping and cleaning the menu information to feed into our chatbot's knowledge base. Teamwork: As a remote team working across different time zones, we gained experience in effective communication, collaboration, and project management. We learned how to leverage tools and platforms to stay connected and productive while working apart.

Challenges We Faced

During the development of Naan Negotiator, we encountered a few significant challenges: Limited Dataset: Training an AI to understand and respond to questions about a specific restaurant's menu was tricky with a small dataset. We had to get creative with our data collection and preprocessing to ensure our chatbot had enough knowledge to be truly helpful. Website Inconsistencies: Scraping menu data from the restaurant's website was complicated by inconsistencies in the site's structure and formatting. We had to develop robust scraping scripts to handle these variations and extract the necessary information. Remote Collaboration: With our team spread across different time zones, coordinating our efforts and staying in sync was sometimes difficult. We had to be proactive in our communication and make use of collaborative tools to keep everyone on the same page. Despite these challenges, our team persevered and learned a great deal in the process. We're proud of what we accomplished with Naan Negotiator and excited to see how it can help make Indian cuisine more approachable and accessible to a wider audience.