2 Sources
[1]
Majority of Indian firms hit by AI agent incident in past 12 months; only 36% contained it in 24 hrs: Survey
A recent survey indicates that Indian organisations are experiencing AI agent security incidents frequently. While 92 percent registered such incidents, only 36 percent managed to contain them within 24 hours. Additionally, the lack of accountability and visibility in AI security governance is
[2]
92% of Indian Organisations Hit by an AI Agent Security Incident, Only 36% Contained It Within 24 Hours: ServiceNow 2026 Report
Leaders describe their agentic security as "established". The controls behind it tell a different story , finds ServiceNow research. 92% of Indian organisations experienced an agentic security incident in the past year, yet only 36% of those organisations contained it within 24 hours, according to
Share
Copy Link
Nine in ten Indian organisations reported a security incident involving AI agents in the past 12 months, and fewer than four in ten contained it within a day. That is the headline finding of a ServiceNow survey of 500 Indian security leaders, published under the title Deployed but Ungoverned: The State of Agentic Security in India.
1
2

Source: CXOToday
ServiceNow, an enterprise software vendor that commissioned the survey through its India and South Asia operation, says it wanted to examine how well Indian organisations can see and govern their own AI agents as deployment grows.
2
The findings land on receptive ground: India's cyber emergency response agency issued an advisory on AI agents in October 2026 that put agent visibility and control on organisations' agendas.2
Only 19% of organisations have a single named person accountable for agentic AI security.
1
Only 18% have unified visibility into the agents they run.1
Just 30% enforce an automated check that blocks agents from deploying until they pass security review.2
Per the survey, only 47% can produce a full evidence package on request, which matters when auditors or incident responders come asking.2
Per the survey, 37% of organisations over the past year found AI agents running in production that their security teams had never registered or did not know about, the report's "shadow AI" blind spot.
2
Some 51% of organisations can automatically spot an AI agent behaving oddly, but 67% cannot automatically act on it.
1
A flagged agent keeps running until a person steps in. The exposure reaches beyond office IT. Per the survey, 83% of organisations run agents in or connected to operational technology and physical environments.2
Only 29% have a separate governance framework covering them.2
Half of the surveyed leaders describe their agentic security as established, and the report itself frames that as confidence running ahead of the controls on the ground.
1
The gap has a compliance edge: the agency's 2022 directions require cyber incidents to be reported within six hours of awareness, so slow containment is not just a security problem.2
The survey also carries a prescription that happens to match what ServiceNow sells. Some 43% of organisations run agentic security on a unified platform, and they report higher automation rates than the rest.
2
Among that group, 62% have automated real-time alerts, against 43% of the others.2
And 45% automatically fix policy violations, against 24%.2
51% capture evidence continuously, against 43%.2
That is a correlation in vendor-commissioned data, not proof that the platform causes the outcome.Kulmeet Bawa, ServiceNow's group vice president and managing director for India and South Asia, puts it: "You can't secure or govern what you can't see."
2
The numbers are self-reported, gathered by a company that sells the kind of platform it recommends, so they measure perception as much as practice.
1
2
Related Stories
Summarized by
Navi
16 Oct 2026•Policy and Regulation

08 Sept 2026•Business and Economy

08 Sept 2026•Technology
