Rob Standing, Regional Vice President for the Middle East, Turkey, and Africa at Rubrik, warns that AI agents are outpacing enterprise security controls, with the Middle East and Africa experiencing rapid adoption, limited visibility, and widening governance gaps amid the expansion of non-human identities.
How are organisations in the Middle East and Africa experiencing the gap between rapidly advancing AI agents and lagging enterprise security guardrails, and what factors make the region particularly vulnerable or resilient?
Organizations across the Middle East are rapidly accelerating AI adoption, supported by strong national strategies and large-scale digital transformation initiatives. For example, the UAE announced a framework in April 2026 to deploy agentic AI across 50% of government sectors, services, and operations within two years. That momentum creates significant opportunity, but it also reinforces the need for strong recovery readiness, comprehensive observability, and security controls to keep pace with this deployment. According to the latest Rubrik Zero Labs (RZL) report, 86% of IT and security leaders expect AI agents to outpace security guardrails within the next year, while only 23% report complete oversight of the agents active in their environments.
With only 23% of global leaders claiming full visibility into AI agents — and many likely overestimating — how should regional CISOs rethink identity governance as non-human identities proliferate across their environments?
CISOs in the region need to rethink identity governance as agentic systems expand, particularly given the rapid growth of non-human identities that are difficult to track and control. To mitigate these risks, organizations must adopt an approach of cyber resilience that has the ability to analyze agentic AI risks across three distinct layers: the tool layer, the cognitive layer, and the identity layer. Our Rubrik research indicates a profound disconnect between perceived control and operational reality, with 80% of leaders claiming strong observability, while 86% anticipate that agentic proliferation will outpace security guardrails within the next year.
Your research highlights the rise of a “shadow workforce” of autonomous agents operating with persistent access. What practical steps should enterprises in regulated sectors like government, finance and energy take to regain control?
Many enterprises still lack visibility into what agents are doing, what systems they access, and whether actions can be reconstructed after an attack. Regaining control, therefore, requires stronger governance across the Identity Layer, with clearer boundaries around agent permissions and access to critical environments. Organizations also need more structured telemetry and auditability, so agent-driven actions are traceable and recoverable, particularly as autonomous systems become more embedded in day-to-day operations.
More than 80% of respondents say AI agents require more manual oversight than they save. How can organisations shift from this operational burden toward a model where AI actually reduces complexity rather than amplifying it?
The operational burden associated with AI is largely driven by gaps in visibility and governance. To ensure AI reduces rather than increases complexity, organizations need stronger identity controls and clearer operational boundaries for autonomous systems. At the same time, resilience strategies must evolve to provide more dynamic oversight and phased recovery models, with granular control over the autonomous workforce.
Nearly nine-in-ten leaders fear they cannot meet recovery objectives as agent-driven threats accelerate. How is Rubrik helping enterprises build resilience when autonomous systems can take actions that are difficult to reverse?
Rubrik research shows that 88% of leaders express concern about meeting recovery time objectives as agentic threats increase, and 33% believe recovery from agentic attacks will lag behind traditional incidents. Rubrik’s approach focuses on improving visibility and traceability so organizations can better understand agent actions, contain incidents, and support more controlled recovery.
As adversaries begin using agentic systems to compress attack timelines and scale operations, what new attack vectors are emerging across the identity, cognitive and tool layers — and how should regional security teams prepare?
As adversaries begin leveraging agentic systems, new attack vectors are emerging. Three critical layers are at risk: the identity layer (token theft and impersonation), the tool layer (unsafe execution and sandbox escapes), and the cognitive layer (prompt injection and logic manipulation). To mitigate any threats, organizations must deploy a defense-in-depth strategy that integrates identity governance, rigorous tool security, and isolated environments.
Your report states that AI strategy is now inseparable from the resilience strategy. What governance, verification and visibility frameworks should boards in the Middle East prioritise to ensure safe, controlled AI adoption at scale?
To ensure safe AI adoption at scale, boards must adopt a layered governance model that treats AI strategy as a core component of organizational cyber resilience. This requires implementing the NIST AI RMF for risk discipline and ISO/IEC 42001 to turn governance intentions into verifiable evidence. In the Middle East, organizations should also prioritize alignment with the NCA Essential Cybersecurity Controls (ECC 2:2024) in Saudi Arabia and the federal PDPL alongside Digital Dubai’s AI Ethics Guidelines in the UAE. Verification must involve testing every use case against failure modes, such as excessive agency, before human sign-off, while visibility requires a live inventory of all agents and identities. Ultimately, if a board cannot prove how to contain or stop an autonomous agent, they have not deployed AI at scale but rather are risking it at scale.