Personalize Travel Planning with AI & Real-time POI Search
Streamline travel planning by providing instant, personalized recommendations based on real-time data, reducing research time by up to 90%.
Planning personalized trips manually with up-to-date information is time-consuming and often leads to generic recommendations. This n8n workflow leverages AI with conversational memory and real-time point-of-interest data to deliver instant, tailored travel advice.
Documentation
AI-Powered Travel Assistant with MongoDB Atlas
This n8n workflow creates a sophisticated AI travel assistant capable of providing personalized recommendations, remembering past conversations, and accessing a real-time database of points of interest. It's ideal for enhancing customer service, internal planning tools, or personal travel discovery.
Key Features
- Intelligent Trip Planning: Leverages Google Gemini to understand complex travel queries and generate tailored suggestions.
- Persistent Conversational Memory: Utilizes MongoDB Chat Memory to maintain context across interactions, offering a seamless user experience.
- Real-time Point-of-Interest Search: Integrates with MongoDB Atlas Vector Search to provide up-to-date, relevant recommendations from your custom knowledge base.
- Effortless Data Ingestion: Includes a dedicated webhook to easily add and update new points of interest, ensuring your AI agent always has the latest information.
How It Works
The workflow operates in two main parts: an interactive AI agent and a data ingestion pipeline.
AI Travel Agent: When a user sends a chat message, the 'When chat message received' node triggers the 'AI Traveling Planner Agent'. This agent, powered by the 'Google Gemini Chat Model', uses the 'MongoDB Chat Memory' to recall previous conversation context. For specific location-based queries, it intelligently queries the 'MongoDB Atlas Vector Store' (acting as a tool) to retrieve relevant points of interest. The 'Embeddings OpenAI' node facilitates vector search by converting queries into embeddings. The agent then synthesizes this information to provide a comprehensive and personalized response to the user.
Point of Interest Data Ingestion: A separate 'Webhook' endpoint ('/ingestData') allows you to easily add new points of interest to your database. Upon receiving data, the 'Default Data Loader' and 'Recursive Character Text Splitter' prepare the text. The 'Embeddings OpenAI1' node generates vector embeddings for this data, which are then stored in the 'MongoDB Atlas Vector Store1' (set to 'insert' mode) within your 'points_of_interest' collection. This ensures your AI agent's knowledge base is continually updated.