Data first, AI second: why the smartest companies start with structure
AI might grab the spotlight, but data runs the show. Without structure, even the most advanced models are guessing rather than learning. Canon’s Norihiro (Nick) Katagiri made that point clearly in a recent iTNews Asia interview, reminding businesses that successful AI depends on solid, well-organised data from the start.
Across Asia, companies are racing to embed AI into their operations. Yet all too often, projects stall because the underlying data is fragmented or incomplete. Gartner expects most AI initiatives built on weak data foundations to be abandoned. The real problem is rarely the technology; it is the lack of structure beneath it.
Katagiri describes this as “data readiness.” It begins with digitising information, then classifying and organising it so that machines can interpret it accurately. It also means investing in people who understand how to use and question what the systems produce. When that foundation is strong, AI becomes far more effective.
It’s a message that resonates far beyond AI. For those of us working with digital reporting, the same principle applies. Structured, high-quality data is what makes everything else possible, whether it’s a regulatory filing, a financial report, or a predictive model – and of course, we believe that XBRL is the most powerful way to structure many kinds of corporate data. Data quality isn’t the boring bit before innovation. It is the innovation.
Read the full interview here.
