Businesses are adopting autonomous AI agents faster than they are putting governance frameworks in place, raising concerns over accountability, cybersecurity and operational resilience, according to new research from Optro.
The study found that one in three organisations already use AI in critical resilience workflows, yet agentic AI failures, including loss of control and autonomous decision-making errors, remain among the least-tested business disruption scenarios. Nearly 30 per cent of organisations surveyed said they had never tested for such failures.
The report, titled When AI Leaves the Chat and Enters the Workflow, argues that organisations are moving beyond conversational AI tools and increasingly deploying autonomous systems capable of taking actions within core business processes. As a result, the challenge has shifted from trusting AI-generated outputs to ensuring accountability for AI-driven decisions.
According to the research, 58 per cent of business leaders believe their governance controls are keeping pace with AI adoption. However, only 18 per cent have active risk mitigation measures in place. The consequences are already emerging, with 40 per cent of organisations reporting inaccurate AI outputs in the past year, 27 per cent reporting data breaches and 26 per cent facing regulatory action linked to AI use.
The study also highlighted growing concerns around accountability. Nearly half of security decision-makers identified agentic AI as a major security risk, while almost two-thirds of organisations reported experiencing an AI agent-related incident over the past 12 months, resulting in data exposure, operational disruption or financial losses.
Another challenge is the lack of visibility into how AI agents are being deployed across organisations. While 85 per cent of respondents said AI had been integrated into core operations, only one-quarter reported having comprehensive visibility into employee use of AI systems. The report warns that many autonomous agents are functioning as non-human identities that can access systems and make decisions without being fully inventoried by IT teams.
“The reality today is that agentic AI adoption is fast outpacing governance,” said Guru Sethupathy, GM of AI Governance at Optro. “But to harness its potential responsibly, leaders must recognize that governance models designed for static manual processes cannot keep pace with autonomous systems of action.”
He added that organisations should view governance as an enabler rather than a barrier to innovation. “Redesigning governance isn’t about pulling back on innovation; it’s about building the control structure to give organizations the confidence to scale AI faster and more reliably than the competition.”
The report concludes that businesses face a narrowing window to establish accountability frameworks before regulators, incidents or remediation requirements force action. Its findings are based on four surveys conducted between February 2025 and May 2026 involving audit, risk, compliance and cybersecurity professionals across North America, Europe and the UAE.
Source: Khaleej Times