More than half of Indian CISOs surveyed, 56%, say third-party or vendor AI tools with unclear data handling are among the biggest AI-related data leakage risks. The finding comes from Trellix’s new research, Mind of the CISO: The New Insider.
The worry goes beyond vendors. 58% of Indian CISOs rank non-human insiders, such as AI agents and machine identities, among their top three security risks. Globally, the figure is 42%.
Insider Incidents Are Already Common
98% of Indian organisations had an insider incident linked to AI or autonomous systems in the past 12 months. Insider incidents cost affected organisations US$2.6 million on average. The global average is US$2.4 million.
Half of CISOs agree they need to significantly improve detection of risks from AI tools and non-human identities. 48% expect AI and automation to be critically important in managing insider risk over the next 12 to 24 months. The global figure is 30%.
Regulation also has a role. 34% of Indian CISOs ranked the Digital Personal Data Protection Act, 2023 (DPDP Act) among the top three regulations or programmes they expect to help most with insider risk. Together with concerns about third-party AI, this shows why companies must know how sensitive data is accessed, processed and protected as AI spreads.
Trellix Wise for Private Cloud
Trellix has also launched Trellix Wise for Private Cloud. The company calls it the first on-premises, federated AI core for the agentic SOC. It extends a security fabric across a customer’s existing infrastructure. The aim is to close the detection and response gap without giving up data sovereignty, governance or cost predictability.
The product keeps data under customer control. It deploys across on-premises, air-gapped, hybrid and private cloud environments. It runs fully inside the customer’s environment, so data stays within the security boundary. It is never sent to a public LLM or external cloud.
Pricing is predictable. It replaces token-based models, so organisations can scale AI without rising costs. Results are auditable. They come with confidence levels, plain-language explanations and detailed dossiers on how the AI reached its view. Analysts get full visibility into each decision.
What Trellix and Its Partner Said
Vishal Rao, CEO of Trellix, said AI is not just speeding up attacks. It is commercialising them. He said Trellix aims to meet AI-powered threats with AI-powered defence, on customers terms.
Ganesh Iyer, Managing Director, Trellix India and SAARC, said Indian organisations must use AI to speed up detection while keeping full control of sensitive data. He said customers need confidence about where their data stays, how the AI reaches its conclusions and what it costs to scale.
Suchit Karnik, COO at RAH Infotech Pvt Ltd, said customers teams often have to piece together what different tools tell them. The product connects those signals with threat intelligence to reveal activity that looks harmless alone.
The product builds on nearly 40 years of Trellix threat intelligence and its Trellix Wise AI technology. Trellix will host a virtual webinar, Agentic SOC without Tradeoffs or Compromise, on October 28.
