The 43-point gap: IT claims vs. reality on AI agent ownership

A new survey from Ivanti, a provider of IT asset management and security solutions, reveals a stark disconnect in how organizations track AI agent ownership. The study, which polled 3,900 employees across six countries including the United States, United Kingdom, Germany, France, Australia, and Japan, found that 85% of IT professionals claim every AI agent has a named owner. However, only 42% of those same professionals report that ownership is actually clear and documented. This 43-percentage-point gap suggests that governance frameworks, while nominally in place, are failing to deliver transparency in practice. The finding underscores a systemic issue where organizational confidence in AI oversight does not match operational reality. According to Ivanti's analysis, no existing governance framework was specifically designed to close this gap, leaving enterprises vulnerable to unmanaged risks.

Leaders hide AI usage at nearly double the rate of other employees

The Ivanti research also highlights a troubling trend in shadow AI adoption among senior leadership. Organizational leaders, defined as executives and managers with decision-making authority, are reportedly concealing their AI usage at a rate of 42%, compared to 23% for all other employees. This nearly two-to-one ratio indicates that those in positions of power are more likely to bypass official AI governance channels. Among these leaders who hide their AI use, 52% cite a desire for a 'secret advantage' as their primary motivation. This finding aligns with broader industry observations that competitive pressures are driving executives to experiment with AI tools without formal approval. The practice not only undermines IT governance but also introduces significant security and compliance risks, as unvetted AI agents may access sensitive corporate data. Ivanti's report suggests that traditional top-down governance models are insufficient to address this behavior.

The governance gap: Existing frameworks fail to address agent ownership

The 43-point discrepancy between claimed and actual AI agent ownership is not merely a measurement error but a structural flaw in current governance approaches. Ivanti's analysis indicates that existing frameworks, such as those based on the NIST AI Risk Management Framework or ISO/IEC 42001, were designed for traditional software assets and do not account for the dynamic, autonomous nature of AI agents. Unlike static applications, AI agents can be created, deployed, and modified by individuals without central IT oversight, especially with the rise of low-code and no-code platforms. The survey found that 68% of IT professionals believe their organization lacks the tools to track AI agent ownership effectively. This gap is further exacerbated by the fact that AI agents often operate across multiple departments and cloud environments, making centralized tracking challenging. Without a dedicated governance model that enforces ownership at the point of creation, enterprises risk accumulating a growing portfolio of unmanaged AI agents.

Shadow AI risks: Security, compliance, and competitive exposure

The concealment of AI usage by leaders and the lack of clear ownership for AI agents create a cascade of risks for organizations. Security teams face challenges in monitoring data flows when AI agents operate outside approved channels, potentially exposing sensitive customer or proprietary information. Compliance departments must contend with regulatory requirements, such as the EU AI Act, which mandates transparency in AI deployment. The Ivanti survey found that 61% of organizations have experienced at least one security incident related to unmanaged AI agents in the past year. Additionally, the competitive advantage sought by leaders through hidden AI use may be illusory, as uncoordinated efforts often duplicate work or produce inconsistent results. The report recommends that organizations implement automated discovery tools to identify all AI agents in use and establish clear ownership policies that apply equally to executives and staff. Without such measures, the gap between claimed and actual governance will likely widen as AI adoption accelerates.

Closing the gap: Practical steps for enterprise AI governance

To address the 43-point ownership gap and the high rate of concealed AI use, Ivanti's report outlines several actionable recommendations for enterprises. First, organizations should deploy automated discovery tools that can continuously scan for AI agents across all environments, including cloud, on-premises, and edge deployments. Second, governance policies must be enforced at the point of AI agent creation, requiring developers and users to register the agent and assign a named owner before deployment. Third, training programs should target leaders specifically, emphasizing the risks of shadow AI and the value of transparency. The survey indicates that organizations with automated discovery tools are 2.5 times more likely to have clear ownership documentation. Finally, Ivanti suggests that enterprises adopt a 'zero-trust' approach to AI agents, treating every unregistered agent as a potential threat until verified. These steps, while requiring upfront investment, can significantly reduce the risks associated with unmanaged AI and improve overall governance posture.