AI property search across 84,000 parcels in Norway

Finding attractive property opportunities was a slow, manual process. Byggkon's team had to search public property systems, evaluate land and ownership data, identify potential sellers, enrich contact information externally, and manage outreach in separate tools. That meant operational overhead, inconsistent evaluations, and a process that did not scale beyond the people who already knew it.

Byggkon wanted one operational platform: AI-assisted property discovery, structured filtering and scoring, integrated lead workflows, and an interface that made complex property data easier to work with day to day.

Designed and built a custom multi-service platform that combines cadastral data, AI-assisted search, lead enrichment, and automated outreach.

Conversational search lets users describe what they are looking for in natural language, for example "residential plots near Stavanger with low terrain slope, long-term private ownership, and development potential". The platform translates that intent into structured queries across the property datasets. Semantic search runs on a vector database so the system ranks by concept and context, not only strict filter matches.

Each property is scored on plot size, ownership duration, number of buildings, elevation, terrain suitability, distance to water, residential zoning, infrastructure access, and the likelihood of seller interest. Direct integration with Norway's Matrikkel services pulls cadastral information, parcel data, boundaries, addresses, and ownership structures.

Once a property is identified, the platform enriches owner contact information, organises leads, generates SMS outreach, and tracks communication history in the same interface. Humans stay in the loop on the decisions that matter; the automation handles the steps that used to bleed time across many tools.

Stack: FastAPI and Next.js for the application layer, PostgreSQL for relational data, Qdrant for vector search, Redis for caching, n8n for workflow automation. Production-ready hosting, queues, and operational monitoring on top.

The platform indexes more than 84,000 properties in the Rogaland region today, with infrastructure designed to extend across Norway over time. Byggkon's team replaces a fragmented manual workflow with one operational system. The platform is in internal production use, with plans to open it up as a commercial product for real estate professionals and adjacent industries.