The Pulse

IBM survey finds gap between confidence and proof of digital control

IBM’s survey found that 94% of enterprises faced a digital sovereignty challenge in the past three years. AI-related problems were the most commonly reported challenge, while 26% of surveyed enterprises reached the index’s leading evidence tier.

AI.info Team · · fact-checked · against 2 sources ↓ · Published 11 Nov 2026

AI-related problems were the most frequently reported digital sovereignty challenge in a survey of large enterprises, cited by 44% of respondents, according to a new IBM index. The study found that 94% of surveyed enterprises had faced at least one sovereignty challenge in the previous three years.

The index assesses whether organizations can retain and demonstrate control over critical AI, technology, operations and data. Among those affected by a challenge, 81% had to provide formal evidence of control to regulators, customers, auditors or internal stakeholders.

IBM published the index on November 11. The research program, developed with Edelman Intelligence, ran from July 17 to August 11, 2026. It included a survey of 2,003 senior decision-makers at large enterprises in the United States, United Kingdom, Germany, France, Saudi Arabia, India and Japan.

Confidence and evidence

An average of 82% of respondents were very or extremely confident that they could provide evidence of sovereign control, the study found. Yet 26% of surveyed enterprises reached the index’s Leading Digital Sovereignty tier, which measures more comprehensive, well-integrated evidence of control.

The index evaluates six dimensions: responsibility, traceability, constraint, intervention, exit and legitimacy. They cover whether an organization can show who was accountable, what happened and whether controls worked, as well as how it responded and recovered. The assessment also considers whether decisions can be justified under scrutiny and whether an organization could change providers or recover operations if needed.

Among all respondents, 33% tested their ability to demonstrate sovereignty continuously or as part of regular governance and assurance processes, according to the findings. The reported rate was 63% among leading enterprises, compared with 29% of those still establishing evidence and 9% at the earliest stage.

Among enterprises in the leading tier, 52% resolved their most testing sovereignty challenge without material disruption, compared with 22% of establishing enterprises, the index found.

AI and other challenges

The AI-related challenges reported by respondents included incorrect outputs, explainability problems, model or provider failures and governance concerns. Other challenges included formal audit or regulatory events, reported by 38% of respondents, and significant provider changes, reported by 35%.

Data-access or transfer issues were cited by 33%, while 30% reported a government restriction. The government restrictions included export controls, sanctions or foreign-jurisdiction issues affecting access to technology, AI models, data, providers or operations.

The research also asked about infrastructure outages, vendor disruption, cyber incidents and regulatory conflicts as sovereignty challenges. The index considers whether organizations can show control when conditions change or dependencies come under pressure.

Collecting proof

The most frequently reported barrier to providing evidence was manual or time-consuming collection, cited by 56%, according to the study. Fragmented evidence across teams and systems, dependence on third parties for records, and legacy systems or inconsistent standards were each reported by 53%. Untested controls, recovery or exit arrangements were cited by 50%.

Besides the survey, the research drew on a review of 27 sources, including legislation, standards, governance frameworks and academic research. The program also included 25 interviews with policymakers, regulators, public-sector officials and policy experts, and perspectives from 17 senior customer leaders across five countries.