Title: Agent-First API for Real Estate Data | Brokk & Sindre Description: Realview wanted to test how AI agents could work directly with its real estate data through a modern, structured interface. Canonical: https://brokk-sindre.dk/en/cases/realview-api - [Home](/en) - [Cases](/en/cases) - Realview api # Agent-First API for Real Estate Data Realview wanted to test how AI agents could work directly with its real estate data through a modern, structured interface. By Mikkel Krogsholm · April 15, 2024 Data creates new possibilities only when systems provide safe, usable access. Agent-ready infrastructure is a leadership choice long before it becomes an API choice. - Fully functional API + MCP server in 2 days - AI agents can run advanced analyses directly on housing data - A working interface delivered as a focused two-day build ## Background Realview’s real estate data is used by banks, mortgage companies and real estate agents for valuation, credit assessment and market analysis. The company had an existing customer API and wanted to test an interface that AI agents could query and analyze through. Realview wanted to test what happens when you give agents direct access to real estate data. And they needed a modern API to build it on. ## What I built I built an API in [FastAPI](https://fastapi.tiangolo.com) with MCP (Model Context Protocol) included in the original architecture. ### Tech stack - **FastAPI** for the API itself - **MCP** (Model Context Protocol) for agent access - **Realview’s database** as the data source It took two days to produce a working API with endpoints, documentation and an MCP server that AI agents can use. ## Results After delivery, we tested the MCP server by having AI agents run analyses and reports directly on Realview’s housing data. The agents retrieved and processed data, then delivered results for review. The work showed that access to data is part of the leadership decision about where AI can create value. ## Related cases ### Danish Data for AI: MCP Servers Danish public registries such as Statistics Denmark and CVR require complex API calls and return large responses when AI systems use them. ### AI-Powered Concert Company How much of a real concert company can specialized AI agents reliably operate across booking, contracts, marketing, communication and payments? ### AI-Powered Risk Assessment for Smukfest Hundreds of artists, thousands of guests, and a security team that needs to prepare for everything from mosh pits to quiet jazz sessions. ## Do you face a consequential AI decision? Write about the decision, or ask about a leadership briefing or talk.