A majority of technology leaders surveyed—52%—say the CIO is held accountable when an AI agent makes an error. CIOs are identified as the party most likely to shoulder responsibility, ahead of functions including customer service, legal, and compliance.

That's a topline finding of the Communications Reckoning: When the AI Agent Fails, Someone Has to Answer survey commissioned by 8x8, Inc. The global survey was conducted in July 2026 among 2,501 CIOs and CTOs. It examined how technology leaders are managing the operational, governance, and organizational challenges associated with the growing use of AI agents.

CIOs Have Become the Last Line of Defense

The survey indicates that responsibility for AI-related failures is frequently concentrated within the technology organization, even when CIOs may not have direct control over the systems, vendors, or data involved.

Among respondents, 52% said the CIO would be held accountable if an AI agent made an error, compared with 16% who identified customer service leadership and 6% who identified legal or compliance teams.

The results highlight a potential gap between accountability and control. As organizations deploy AI across multiple applications, vendors, and business functions, technology leaders may be expected to manage outcomes without having complete visibility into how AI systems operate or access to all of the information needed to investigate incidents.

Data Sovereignty is Becoming a Board-Level Consideration

Data location and sovereignty are also emerging as significant considerations in AI infrastructure and vendor decisions.

Overall, 82% of respondents said the location of AI infrastructure matters to their organizations. Thirty percent (30%) identified it as the primary factor in vendor decisions, while another 52% described it as one of several critical considerations.

The findings point to growing concern about where organizational data is stored, processed, and transferred, particularly as businesses operate across jurisdictions with different regulatory requirements. Regulations such as GDPR, HIPAA, CCPA, and comparable data-protection frameworks can create additional requirements around data handling, residency, and cross-border transfers.

The issue is not limited to organizations operating in Europe. Technology leaders in the United States and other markets are also increasingly considering data sovereignty as part of their AI strategy and risk-management processes.

Vendor Consolidation Remains Difficult

Organizations continue to face challenges in simplifying increasingly complex technology environments. However, the research suggests that a lack of capable technology is not the primary barrier to consolidation.

Only 9% of CIOs said that no single platform can meet their organization's requirements. Instead, respondents most frequently cited organizational, financial, regulatory, and commercial barriers:

• 30% cited migration cost and complexity.
• 24% cited regulatory or data-residency constraints.
• 16% cited vendor lock-in.
• 15% cited internal politics and stakeholder alignment across teams.

These findings suggest that technology consolidation is often constrained less by functionality than by the practical challenges of migrating existing systems, meeting regulatory requirements, managing contractual dependencies, and aligning stakeholders.

Governance Remains a Significant Gap

As AI becomes embedded in more business processes, communications and other technology infrastructure increasingly intersect with questions of accountability, data management, security, and vendor oversight.

For CIOs and other technology leaders, the challenge is therefore not simply selecting AI-enabled tools. It also involves establishing sufficient visibility across the systems already in use, understanding where data is processed, determining who is responsible when AI systems fail, and managing the risks created by increasingly fragmented technology environments.

The research suggests that organizations may need to align responsibility for AI outcomes with the authority, visibility, and governance capabilities required to manage those outcomes effectively.

The findings are part of a broader research series examining AI complexity, governance, and the challenges created by multi-vendor technology environments. Further research is expected to explore these issues throughout 2026 and 2027.

Source: 8x8, Inc.

Stay Connected with ISE Magazine 

Subscribe to our newsletters and magazine for the latest telecom insights, explore the current issue for in-depth features and strategies, and register for upcoming webinars to learn directly from industry leaders.