NTT Docomo exec: AI fails on data, not models

The discussion, between NTT Docomo executive officer and information systems head Shusaku Ijiri and Everpure Japan chairman Ichiro Nakagawa, starts from a shared observation: many companies have run AI pilots successfully but cannot take them company-wide. Ijiri attributes this to the way data is organized. The same customer ID can mean different things to sales, contracts, support and service operations, and AI cannot tell those meanings apart on its own.

That leads both speakers to argue that the old pattern — each application holding its own data — no longer fits. Everpure introduced the idea it calls "Data Primacy" at its annual "Pure Accelerate" event held in Las Vegas in June 2026, which shifts the focus from applications to data and builds governance into the data layer. Ijiri describes the same direction as "AI Ready Data": not data merely reshaped for AI, but a corporate asset that stays reusable as uses change.

Where the two land is that continuity and control matter as much as capability. Ijiri notes that a data platform outage would hit customer service, network operations and decision-making, and that understanding data for AI use and identifying critical data for emergencies are the same work — AI use, data governance and business continuity planning are not separate topics. Whether companies can hold on to that control as AI and cloud services change is likely to shape who turns pilots into real change.