Why AI needs XBRL
By John Turner, CEO of XBRL International, based on an original post on LinkedIn.
XBRL is outdated? That’s news to AI—and suggestions from some preparers that AI can replace XBRL remain wildly optimistic!
Reality check: AI without structured data is like a self-driving car without roads, maps and experience on the road. It might get somewhere, but not where you want to go.
XBRL isn’t a burden; it’s the fuel for AI-powered analysis, ensuring accuracy, comparability, and decision-usefulness. Ditch structured data in favour of AI scraping PDFs? Sure—if you want your investment models built on guesses and markets running on vibes.
The only certainty with an AI “magic wand” is that every model will generate a slightly different version of the truth. Investors don’t need hallucinations—they need reliable, traceable numbers and commentary directly from companies, signed off by management.
And about that “huge cost” of digital reporting? Companies already sink thousands—often tens of thousands—of hours into annual reports: data collection, consolidation, review, sign-offs, assurance, polished design. The tagging? A few dozen person-hours. A tiny investment for a massive payoff: one definitive, human-and-machine-readable version of the truth that management stands behind. There is more software on the market than ever before, so do shop around. More training? More clarity around assurance? Sure, that all makes sense.
Fundamentally, digital reporting means that regulators and investors reap the rewards, but so do issuers: better benchmarking, greater visibility, smoother access to capital. And as digital-first reporting evolves, streamlined systems make compliance easier—not harder.
Oh, and XBRL isn’t “old tech.” It’s evolving constantly, modernised for today’s digital reporting needs. High-reward, low-effort, and indispensable for decision-making.
Structured data isn’t the problem—it’s the solution.
