SIAgentNews.com
← All news
DevelopingAgentsSecurity

Most Indian firms face AI agent security incidents with slow response times

A new survey reveals that while 92% of Indian organizations have experienced AI agent security issues, fewer than half can contain them quickly.

· First reported by livemint.com

92 per centIndian organizations hit by AI agent security incidents

The short version

  • 92% of Indian organizations reported an AI agent security incident in the past 12 months.
  • Only 36% of firms were able to contain these security incidents within 24 hours.
  • 37% of organizations found AI agents running in production that were never registered with security teams.
  • 67% of firms cannot automatically act when an AI agent behaves in an anomalous way.
  • Only 18% of surveyed companies have a unified, single-interface view of all their agents.

Why it matters

As companies integrate AI agents into their daily operations, the lack of centralized visibility and automated response tools creates significant cyber risks. Organizations need to move from simply deploying these tools to establishing formal governance and automated security gates.

The scale of AI security incidents

A recent survey of 500 security leaders conducted by ServiceNow highlights a significant gap between AI adoption and security readiness in India. The data shows that 92 per cent of Indian organizations experienced a security incident involving an AI agent over the past year. This high frequency suggests that AI agents are becoming a standard part of the corporate landscape, but they are also introducing new vulnerabilities.

Despite the widespread nature of these incidents, the speed of response remains a major concern for many businesses. The report found that only 36 per cent of the organizations surveyed were able to contain a security incident within 24 hours. This delay can allow threats to persist longer, potentially leading to more extensive data breaches or operational disruptions.

The findings indicate that while companies are eager to use AI to improve efficiency, they may not be prepared for the risks associated with autonomous systems. The report, titled "Deployed but Ungoverned: The State of Agentic Security in India," suggests that the infrastructure for protecting these agents is currently uneven. Many firms are operating in a reactive mode rather than having proactive defenses in place.

Gaps in governance and visibility

One of the primary issues identified in the study is the lack of clear accountability and oversight for AI agents. Only 19 per cent of the surveyed organizations have a single named individual who is responsible for the governance of agentic AI security. Without a clear owner, it is difficult for companies to establish consistent policies or ensure that security standards are being met across different departments.

Visibility into these systems is also remarkably low, with only 18 per cent of companies having a unified, single-interface view of all their agents. This lack of a central dashboard means that security teams may not know exactly which agents are active or what permissions they have. This obscurity makes it harder to monitor the behavior of AI tools that are interacting with sensitive company data.

The survey also revealed that 37 per cent of organizations discovered AI agents running in production that had never been registered or known to the security teams. These "shadow" AI agents represent a significant risk because they bypass the standard security checks that are meant to protect the network. When agents operate outside of official oversight, they can create unpredictable entry points for cyberattacks.

Limitations in automated defense

While many companies are trying to automate their security, the current tools often fall short of providing real-time protection. The report notes that 51 per cent of organizations have automated real-time detection for anomalous AI agent behavior. However, this detection does not always lead to immediate action, as 67 per cent of firms cannot automatically act when an agent begins to behave in an unusual way.

To bridge this gap, security experts argue that systems need to move at "machine speed" to be effective against modern threats. Kulmeet Bawa, GVP & Managing Director for India & SAARC at ServiceNow, noted that knowing what an agent can reach is not the same as seeing what it is doing. This distinction highlights the need for continuous oversight that monitors the actual actions of an agent rather than just its theoretical capabilities.

Furthermore, only 30 per cent of organizations enforce an automated deployment gate that blocks agents until they pass specific security checks. Without these gates, agents can be introduced into the environment without a thorough vetting process. This lack of automated enforcement makes it difficult for security teams to keep pace with the rapid deployment of new AI technologies across the enterprise.

The path toward cyber resilience

The report emphasizes that for India to achieve its AI ambitions, it must develop a robust framework for understanding and governing the agents it deploys. This involves moving beyond simple deployment toward a model where every agent is registered, owned, and visible in one central location. Establishing these foundations is necessary to ensure that AI innovation is supported by a layer of trust and cyber resilience.

Organizations are encouraged to focus on creating complete evidence packages, which 47 per cent of surveyed firms can currently produce on demand. Having a clear audit trail is essential for identifying how an incident occurred and what data may have been compromised. Strengthening this capability allows for faster forensic analysis and more effective long-term security planning.

Ultimately, the goal for Indian enterprises is to match their technological growth with sophisticated security controls. By implementing automated controls and continuous oversight, companies can mitigate the risks of autonomous agents. This balanced approach ensures that AI remains a tool for productivity rather than a source of significant security liability for the organization.

What to watch

  • Whether Indian regulatory bodies introduce specific security standards for AI agents.
  • The adoption rate of automated 'deployment gates' to block unvetted AI tools.
  • How many more 'shadow' AI agents are discovered as companies increase their AI usage.

Try it

Ask your IT or security team if there is a central registry or a single dashboard where all company-approved AI agents are listed and monitored.

Sources and links

First reported by livemint.com. Written by the SI Agent News Desk from the reporting above.Written with AI from the source above. Every number is checked against the source. Spot a mistake? Corrections.

Liked it? Say why

Add a line for other readers.

No account needed. We never show your email. Comments with links, ads or abuse are held and not shown. Your comment shows after a quick check.

More AI news