
India AI Security Gaps Exposed As Enterprises Deploy Agents Unchecked
Illustration: NexVani Media
A new study reveals that 92 percent of Indian organisations faced an AI security incident last year, highlighting growing gaps in governance and control.
Indian organisations are rapidly adopting artificial intelligence, but their ability to secure and govern these systems is falling behind. A recent report by ServiceNow reveals that 92 percent of organisations in the country experienced an AI agent security incident over the past year. Despite high confidence among leadership, only 36 percent of affected entities managed to contain an incident within a 24-hour window.
The findings from the report, titled Deployed but Ungoverned: The State of Agentic Security in India, indicate a widening disparity between the rollout of AI tools and the implementation of proper security frameworks. While half of the surveyed security leaders described their agentic security as established, only 19 percent reported that a single named individual was accountable for governance, and just 18 percent maintained unified visibility into their operations.
Compliance and deployment controls remain weak across many firms. Only 30 percent utilize automated deployment gates to block unverified AI agents, and less than half can produce a complete evidence package on demand. Furthermore, shadow AI presents a significant blind spot, with 37 percent of organisations discovering unregistered agents running in production environments over the past twelve months. Security teams are frequently brought in too late, with only 46 percent involved prior to development.
Unchecked AI agent adoption leaves many Indian organisations struggling to contain security incidents swiftly.
Operational technology and physical environments face elevated risks as well, with 83 percent of firms deploying AI agents in these settings while only 29 percent maintain dedicated governance frameworks. Although 51 percent of organisations utilize automated real-time detection for anomalous behaviour, 67 percent cannot automatically remediate threats without human intervention.
Regulators are increasingly focusing on these vulnerabilities, supported by guidelines such as CERT-In's advisory on AI agents and mandatory incident reporting frameworks. Organisations leveraging unified platforms demonstrated superior capabilities in automated alerts and remediation, emphasising the need for centralized oversight as autonomous systems scale.
Why it matters As the adoption of autonomous artificial intelligence accelerates across enterprises, the inability to govern and secure these systems introduces severe operational vulnerabilities. For businesses and technology leaders in India, closing the visibility gap and ensuring rapid incident containment are critical to preventing major security breaches.
Sources
More like this

Paris Hilton and AOC Speak Out Against AI Deepfakes

British Hacker Sentenced to Six Years in US Prison Over Crypto Theft Scheme

Major Study To Explore Social Media Ban

