Structured data analysis – with a simple conversation?
One of the big possibilities we’ve been talking about as the AI revolution unfolds is the prospect of interacting with XBRL data – and getting accurate, decision-useful insights – using AI chatbots, without the need for in-depth technical analytics skills. That future is already here.
In case you missed it earlier in the year (we did!), don’t wait to catch up on Stefano Amorelli’s blog and series of video demos showcasing his open-source SEC EDGAR MCP. This leverages the Model Context Protocol (MCP), an open standard for integrating AI-powered tools with external data. The US Securities and Exchange Commission (SEC) provides an incredibly rich dataset of XBRL filings through its EDGAR portal. This tool makes it possible to use an AI chatbot of your choice to interact with the data via the EDGAR XBRL APIs – using simple, natural-language questions to get sophisticated answers.
In this video, for example, Claude accesses SEC XBRL data and creates a detailed breakdown and visualisation of Apple’s financial assets. Tools like this promise to extend the benefits of digital reporting to many more users. They provide access to authoritative, unambiguous, traceable XBRL data and provide a solid foundation for trustable AI analysis – avoiding the need for AIs to interpret documents, scrape data and make guesses about report contents.
And for the regulators among our readers? It’s becoming increasingly clear that machine-readable APIs are an essential component of data access, visibility and machine understandability – all essential in optimising AI outcomes. Not only is it best practice to ensure free access to XBRL data via a central portal, it’s also important that that data is easy to consume – for both AIs and more traditional automated software tools. That means providing APIs should be a serious and urgent consideration.
