Using AI to analyse XBRL reports?
If you missed our latest member-exclusive webinar on 20 November, don’t worry: the recording is now available! The 30-minute session explored how Large Language Models (LLMs) can be used to analyse XBRL reports, from summarising key figures to extracting insights from narrative disclosures.
Presented by our own Revathy Ramanan and Catalina Ibáñez Gutiérrez (Lucanet and Chair of the Implementation Awareness Task Force), the session walked through practical steps for converting real XBRL reports into LLM-ready formats like xBRL-JSON. Participants saw live demonstrations of how to upload, query, and interpret structured XBRL data, both numerical and narrative, using an LLM interface.
While not production-ready just yet, this experimental workflow showcased the promise of AI-powered XBRL analysis, giving us something to look forward to in the world of AI. The session also covered best practices, transparency issues, and important caveats around reasoning and data integrity.
As LLMs continue to evolve, understanding their potential, their limitations, and how they interact with XBRL data is increasingly important for software developers, regulators, and reporting professionals. The future of AI-assisted financial reporting and analysis could be closer than you think.
Catch up on the recording here.
