AI · 18 October 2026
AI Agent Security Incidents: 36% of IT Leaders Would Stay Silent
A Workato report covered by CIO Dive finds 36% of IT leaders would try to handle an AI agent-driven security incident quietly, without disclosure, as shadow AI rises.
What happened
Around one in three IT leaders, specifically 36%, say they would try to handle an AI agent-driven security incident quietly without disclosing it, according to a Workato report covered by CIO Dive. The finding comes as shadow AI, meaning AI tools used outside formal IT oversight, is rising inside organisations.
The research highlights a tension between how some leaders expect to respond when autonomous AI agents cause security problems and how those incidents would be surfaced if transparency were the default. CIO Dive presents the findings as part of a wider picture of IT executives preparing for agent-related security events.
Why it matters
AI agents are moving from experimentation into workflows where they can act on systems, data and customer requests. When an agent fails or is compromised, the consequences reach the people it serves as well as the security team. A decision to keep an incident quiet is therefore a decision about trust, not just about internal reporting lines.
For leaders in experience, AI and digital transformation, the report suggests that governance for agents needs to cover visibility as well as controls. If teams cannot see which agents are running, or feel pressure not to report problems, the organisation loses the information it needs to fix issues before customers notice them.
The Renascence take
Most coverage will treat this as a security readiness story. We see a behavioural one. People rarely hide incidents because they are malicious; they hide them because the cost of disclosure feels higher than the cost of silence, and that calculation is shaped by how the organisation has responded to mistakes in the past.
The real risk is not that an agent will fail, because it will. The risk is that the people closest to the failure have been trained, implicitly, to manage it privately. Customer-obsessed operators should make reporting safe and fast, reward early disclosure, and treat every agent incident as a service-recovery event that needs an owner and a customer-facing response plan. Silence is a design choice, and it can be designed out.
Sources
This briefing was written by our Newsdesk, synthesising reporting from the outlets below. Follow the links for the original coverage.
FAQ
Questions we get on this topic
More in AI
Stay ahead of CX
Get the signal, not the noise.
The stories shaping customer experience — plus the Journal and Experience Loom — in your inbox.


