Automate Personalized Movie Recommendations with AI RAG Chatbot
Deliver highly personalized movie recommendations in real-time, reducing manual search efforts and enhancing user satisfaction by over 90%.
Manually searching for movie recommendations often yields generic results and consumes valuable time. This workflow automates personalized movie discovery by leveraging an AI-powered RAG chatbot, providing tailored suggestions instantly.
Documentation
Automate Personalized Movie Recommendations
This workflow builds an intelligent Retrieval-Augmented Generation (RAG) chatbot that provides highly personalized movie recommendations. By integrating Qdrant for efficient vector search and OpenAI for natural language understanding, it automates the process of discovering movies tailored to specific user preferences.
Key Features
- Dynamic AI-driven movie suggestions based on positive and negative user preferences.
- Efficient data ingestion: Automatically loads movie descriptions and metadata into a vector database for semantic search.
- Seamless integration with OpenAI for advanced natural language understanding and embedding generation.
- Scalable vector search powered by Qdrant ensures fast and highly relevant recommendation results.
How It Works
This workflow operates in two main phases: initial data ingestion and real-time chatbot interaction.
- Data Ingestion: A CSV file containing movie data (like descriptions and names) is pulled from GitHub. These descriptions are then processed and transformed into numerical vector embeddings using OpenAI's embedding model. Finally, these embeddings are stored in a Qdrant vector database, creating a searchable knowledge base.
- Chatbot Interaction: When a user sends a chat message, an n8n AI Agent intercepts the request. The agent utilizes an OpenAI chat model (GPT-4o-mini) and conversational memory to understand the user's intent and preferences.
- Recommendation Logic: The AI Agent triggers an internal n8n tool, specifically designed to handle movie recommendations. This tool takes the user's positive and negative movie preference examples, converts them into OpenAI embeddings, and then queries Qdrant's advanced recommendation API to find the most relevant movies based on semantic similarity.
- Personalized Output: The retrieved movie metadata and recommendation scores are formatted by the agent. The top 3 personalized movie suggestions are then presented to the user through the chat interface, without displaying the internal scores.