According to the latest study by IBM Japan, approximately 89% of Japanese companies do not have sufficient visibility into dependencies within their AI environments, highlighting the realities of multi-vendor use and the need to address requirements that vary by region.
A joint study by IBM Institute for Business Value and Oxford Economics analyzed the control and resilience of the full AI stack and its impact on business performance, based on responses from 1,000 business executives across 16 countries and 17 industries. The study found that many companies in Japan do not have a complete understanding of their AI dependencies.
While 74% of respondents globally and 70% in Japan use multiple vendors in their AI environments, this was attributed to independent decision-making by business units and the need to address regional requirements, highlighting the difficulty of exercising cross-functional governance.
In Japan, 89% of executives said they do not have sufficient visibility into their dependencies on the AI vendors, models and infrastructure they use. The global average number of AI-related incidents and disruptions over the past 2 years was 6, with Japan at a similar level.
While 71% of respondents globally and 69% in Japan said switching major vendors or models is difficult, data location and data sovereignty requirements were identified as challenges.