8×8, a business communications software company, said that over half of surveyed technology leaders hold the CIO accountable when an AI agent makes an error. The share was well ahead of customer-service leadership at 16% and legal or compliance teams at 6%.
The findings come from “Communications Reckoning: When the AI Agent Fails, Someone Has to Answer,” a survey 8×8 commissioned through Censuswide in July and published Sept. 8. Censuswide surveyed 2,501 CIOs and CTOs across the U.S., U.K., France, Australia and the Republic of Ireland.
8×8 argues that the numbers expose a gap between accountability and control. CEO Samuel Wilson said CIOs may not have chosen the AI vendor or have access to its audit trail, yet may still be held responsible when something goes wrong.
Accountability meets data control
The survey also connects control to infrastructure decisions. Among responding CIOs, 82% said where AI infrastructure is hosted affects platform-buying decisions. For 30%, location was the primary consideration, while another 52% described it as one of several critical factors.
8×8 linked those concerns to regulatory pressure tied to GDPR, HIPAA and CCPA. The company also pointed to separate research it released in August in which 89% of surveyed business leaders said data sovereignty had moved higher on their priority list over the previous 12 months. Both figures were published by 8×8.
The survey also asked CIOs about consolidating their technology stacks. Only 9% said no single platform could meet their organization’s requirements. The most cited obstacle was migration cost and complexity at 30%, followed by regulatory or data-residency constraints at 24%, vendor lock-in at 16% and internal politics at 15%.
8×8 interprets that result as evidence that technical capability is not the main obstacle to consolidation. That interpretation also overlaps with the company’s commercial position: 8×8 sells an integrated communications and customer-experience platform and argues that bringing communications data together can give organizations a clearer record of AI activity.
Survey measures perceptions, not failures
The figures describe how CIOs and CTOs assign accountability. They do not count AI incidents, establish their causes or show that CIOs were responsible for the decisions that led to individual failures.
8×8 said the research is the first installment in a three-part “Communications Reckoning” series that will continue throughout 2026 and 2027.