Concept2025
CarScout is a vehicle discovery and stock locator web app built as a proof of concept. Users can search through a guided wizard or conversational chat, then refine the result through a full configurator before checking nearby availability.
Next.js
Gemini
Neo4jProduct Flow
Vehicle Discovery
Users can begin with a guided wizard or describe what they want through chat, making the first step into vehicle search faster and more flexible.
Query and Response Architecture
From intent to nearby stock
CarScout uses a graph RAG pipeline backed by Neo4j to move from broad user intent to relevant nearby vehicle stock. An LLM first extracts structured features from the request, enabling an initial pre-filter. From there, semantic search and graph relationships work together to identify vehicles that match the user's requirements and are available in the vicinity.
