79.1% of Japanese firms say data is ready; only 3.7% see no problems

Mer's report organizes the requirements for AI-ready data into three conditions: quality, connection, and maintenance. The survey results show a notable gap in each area. While 79.1% of respondents consider their data organized, the specific problems they cite—such as missing data (54.5%) and inconsistent notation (44.2%)—indicate that self-assessment may overstate actual readiness.

The connection condition is even weaker: only 10.0% can cross-reference nearly all data by customer, and 41.4% say data is siloed by system with no linkages. Without a customer-centric data structure, AI struggles to form a complete view of the customer, which limits its usefulness for personalized or context-aware tasks.

Maintenance practices also lag. 37.0% of respondents have no ongoing efforts to keep data organized, and only 8.3% have automated systems for continuous cleanup. This suggests that even companies that have cleaned their data once may see it degrade over time. Mer argues that firms need to shift focus from which AI to use to what data to feed it, and proposes a three-principle approach: organize, connect, and maintain.