PropTech

Core Event A developer has modernized Spain’s official cadastre (Sede Electrónica del Catastro) API—a legacy SOAP service dating back to 2003—by building "Predio," a modern JSON wrapper. Crucially, the project includes a Model Context Protocol (MCP) server, enabling LLMs and AI Agents to query, interpret, and analyze complex real estate data directly from government sources. ▶ Modernizing Legacy Debt: By wrapping archaic SOAP interfaces into developer-friendly JSON, the project rescues authoritative data from "digital archaeology" and brings it into the GenAI era. ▶ MCP as the Universal Connector: This implementation highlights the Model Context Protocol’s role as the definitive bridge between LLMs (like Claude) and siloed, structured geospatial data. ▶ Vertical SaaS Arbitrage: Modernizing "ugly" government infrastructure presents a massive opportunity for PropTech startups to build high-value services atop previously inaccessible data. Bagua Insight While Silicon Valley obsesses over parameter counts, the real-world utility of AI is often throttled by data silos locked in 20-year-old XML schemas. Spain’s cadastre API is a prime example: the data is authoritative and mission-critical, yet its integration friction is a barrier to entry. The Predio project underscores a fundamental truth: The ceiling of an AI Agent’s utility is defined by its access to legacy infrastructure. By leveraging the MCP protocol, the developer bypasses the need for model-specific plugins. This "wrap once, deploy to any agent" strategy signals a looming wave of "AI Adapters" for regional and industry-specific legacy systems. We are witnessing a massive "soft-refactoring" of global digital infrastructure, where the goal isn't to replace old systems, but to build the necessary plumbing to make them AI-ready. Actionable Advice For Developers: Target high-value, high-friction sectors like GovTech, LegalTech, and FinTech. Building MCP-compliant wrappers for legacy APIs is a high-leverage move in the current Agentic workflow boom. For Enterprise Architects: Don't wait for legacy vendors to modernize their stacks. Implement lightweight JSON/MCP middleware to expose internal data to LLMs with minimal overhead. For Investors: Look for "Data Plumbing" startups that specialize in transforming non-structured or legacy data into AI-ready formats. These tools represent the essential infrastructure for the next phase of enterprise AI adoption.