Biscuit AI - Personalized Wine and Beverage Recommendation Chatbot

Project Objective

The Challenge:

The goal of this project was to create a smart chatbot that helps users find the perfect wine or beverage based on their personal preferences. We aimed to go beyond basic recommendations by considering factors like wine types, tasting profiles, and food pairings. The challenge was to design a system that could understand user needs and provide accurate, personalised suggestions.

Complexity and Innovation:

The complexity of Biscuit AI came from combining various types of information and using advanced technology to deliver personalised recommendations. We innovated by integrating multiple AI tools and techniques to provide not just recommendations but also useful information about wines, making it a versatile and helpful chatbot.

The Process

Client Collaboration:

We started by working closely with our client to understand their vision. They wanted a chatbot that could provide tailored recommendations for wines and beverages based on users’ interests and needs. This included suggesting wines for special occasions and providing detailed information about different wine varieties. We used this feedback to build a solution using advanced AI technologies.

Technology Stack: To build Biscuit AI, we used:

Frameworks: LangGraph, LangSmith, and LangChain , LLM, openai for managing different tools and functions.

Databases: Pinecone DB for storing and retrieving wine information, and MySQL for managing inventory and checking wine availability.

RAG: For retrieving relevant wine information from our knowledge base.

Chatbot Functionality:

Biscuit AI offers more than just wine recommendations. It helps users by:

Personalised Recommendations: Suggesting wines based on user preferences like wine type, tasting profile, and price range.

Food Pairing: Recommending wines that go well with specific meals.

Inventory Management: Checking if the recommended wines are available and providing information about them.

Special Occasion Suggestions: Offering recommendations for special events like birthdays.

Knowledge Sharing: Providing facts and details about different wines and trends.

Privacy and Data Handling: We ensure user privacy by securely storing only the necessary data to improve recommendations. The chatbot handles user information responsibly and follows data privacy standards.

Feature Inventory

• Personalised Beverage Recommendations: Suggests wines based on user preferences and tastes.
• Dynamic Inventory Checks: Ensures recommendations reflect current stock and reviews.
• Food Pairing Assistance: Recommends wines that complement specific meals or cuisines.
• Special Occasion Handling: Offers tailored suggestions for premium or prestigious wines.
• Knowledge Base Insights: Provides informative content about different wine varieties and trends.

The Results

Biscuit AI successfully delivered personalised and useful wine recommendations, meeting users' needs for both everyday choices and special occasions. The project helped us learn more about using AI frameworks and integrating different data sources. Biscuit AI has become a valuable tool for anyone looking for customised wine recommendations, making it easier for users to find their perfect beverage.