Bridging the Gap: AI and Official Business Data
Verifying company information, especially for compliance and due diligence, often involves tedious manual lookups in official registries. This process can be time-consuming and prone to human error. For one developer, performing compliance checks on Spanish suppliers for a client highlighted the inefficiency. The task required manually querying Spain's official business registry, BORME, for each company, extracting key details like status, directors, and financial history, and compiling them into a spreadsheet. What took an afternoon for 40 companies prompted a more ambitious solution: empowering an AI to handle the task directly.
The core idea was to leverage Claude, a powerful large language model, by providing it with direct access to real-time data from official registries. This bypasses the need for manual intermediary steps. The solution involved building a series of Microservice Communication Protocol (MCP) servers. MCP is designed to facilitate communication between different software components, and in this context, it acts as a bridge between Claude and the APIs of national business registries.
By connecting Claude to these MCP servers, the AI can now perform complex queries, such as identifying dissolved companies or those in insolvency proceedings, across a large list of businesses. Claude then summarizes the findings, presenting actionable insights without requiring the user to navigate the intricacies of each registry's interface or data format.
A Multi-National Data Network for Claude
The project has expanded significantly, with the developer creating seven distinct MCP servers to cover various European countries. This distributed network provides Claude with a broad reach into official business data.
- Spain BORME: This server taps into Spain's official business registry, BORME, which is updated daily from the official State Bulletin (BOE) feed. It covers approximately 2.8 million companies, offering real-time data on company status, directors, capital, and corporate actions.
- France SIRENE: Serving France's extensive business landscape, this server connects to the SIRENE database via api.gouv.fr. It provides access to information on an estimated 25 million companies.
- United Kingdom Companies House: While the source text does not detail the specific data points for the UK, it confirms the existence of a dedicated MCP server for the UK's Companies House, a primary source for UK company information.
The initiative aims to build out coverage for other key European registries. The strategic advantage of this approach is the ability to perform cross-border verifications and data aggregation with unprecedented ease. Instead of logging into multiple national portals or relying on potentially outdated third-party databases, a single query to Claude, mediated by these MCP servers, can yield comprehensive results.
The Technical Underpinnings and Future Implications
The architecture relies on the MCP protocol to standardize the communication flow. Each server acts as an adapter, translating Claude's requests into the specific format required by the target registry's API and then transforming the registry's responses back into a format Claude can understand and process. This abstraction layer is crucial for enabling a cohesive user experience, regardless of the underlying registry's technical implementation.
For developers and compliance professionals, this means a dramatic reduction in the time and effort needed for business verification. Tasks that previously involved hours of manual work can now be executed in minutes, allowing for more frequent and thorough checks. The potential applications extend beyond simple compliance; it can aid in market research, competitor analysis, and even detecting fraudulent entities by cross-referencing information across different jurisdictions.
The surprising detail here is not the technical feat of building MCP servers, but the directness with which this enables an LLM to interact with structured, official government data. This moves beyond LLMs merely summarizing publicly available web content to actively querying and interpreting authoritative databases. It’s a step towards AI agents that can perform concrete, data-driven tasks in the real world.
What Lies Ahead?
The current setup focuses on Spain, France, and the UK, but the ambition is to extend this to all EU member states. Building out coverage for the remaining registries will create a truly comprehensive AI-powered tool for business intelligence and compliance across the European Union. The developer's work directly addresses a gap in how businesses can efficiently access and utilize critical official data, transforming a laborious manual process into an AI-driven workflow.
What nobody has addressed yet is the scalability and maintenance burden of keeping these seven (and potentially more) MCP servers synchronized with the constantly evolving APIs and data schemas of national registries. Ensuring data accuracy and uptime across such a distributed system will require robust monitoring and continuous development.
This project demonstrates a practical, impactful application of AI by connecting it to authoritative data sources. It’s a blueprint for how LLMs can become more than just conversational tools, evolving into powerful engines for data retrieval and analysis in professional contexts.