Some Next Generation XBRL thoughts
At Data Amplified 2026 this week we kicked off proceedings with several presentations on Next Generation XBRL, for now known as “Project Tavi”.
Four things in particular seemed to resonate with the audience as we walked through a number of demonstrations:
1. Combining the definitions and relationships contained in XBRL taxonomies (including company-specific definitions and relationships) with the data in XBRL reports in a single (JSON for now) machine-readable format creates an incredibly rich analytic resource, both for AI and traditional analysis tools.
2. Creating a new set of “Cube Constraints” will give regulators and other data collectors far greater control over data quality, including the automation of a host of rules typically contained in today’s Filing Manuals.
3. Not only does Next Generation XBRL provide direct access to data and metadata for analysis, it also gives standards setters and other taxonomy authors a simpler, easier and hugely powerful way to create their data definitions. The potential to create digital models of disclosure or regulatory rules (aka taxonomies) directly from those rules or legislation using an AI collaborator will be attractive to many.
4. The opportunity the new format affords for breakthrough performance improvements could enhance workflow and reduce the bottlenecks that reporting tends to create (when millions of companies all report on the same day, for example!).
The Project Tavi specification is now available as a Public Working Draft from our specifications site. It remains an early indication of the direction of travel and there is much to do, but (thanks to LLM-supported coding) it is absolutely the right time for interested stakeholders to experiment, and to provide feedback. Stay tuned for some of our experiments in the coming weeks!
