AI-Powered Travel Planning and POI Retrieval Workflow

This n8n workflow is designed to create an intelligent travel assistant that integrates language models, vector search, and a robust memory system to provide dynamic travel recommendations and information about points of interest (POIs). The workflow begins with an incoming chat message, which triggers interactions with Google’s Gemini language model and OpenAI embeddings for contextual understanding and vector-based search. It stores conversation history in MongoDB to maintain context and retrieves relevant POIs from a MongoDB Atlas vector store using similarity search. Users can also ingest new POI data via a webhook, which is embedded and indexed for quick retrieval. This setup is ideal for building conversational travel assistants, smart tour guides, or interactive itinerary planners that deliver personalized, up-to-date travel info based on user queries.