Thales is expanding its collaboration with Google Cloud to help businesses secure AI agents as they become increasingly capable of making decisions and taking actions across enterprise systems.
The companies said the integration of Thales’ AI Security Fabric with Google Cloud’s Gemini Enterprise will provide visibility and policy controls across interactions between users, AI agents, models and tools.
The move reflects a growing security challenge as businesses shift from AI assistants, which generally respond to user requests, towards agentic systems that can plan tasks, make decisions and act with less direct human intervention.
“Enterprises are moving from AI assistants to AI agents that can autonomously take action, make decisions, and interact with critical business systems,” said Eva Rudin, Senior Vice-President of Cybersecurity Products at Thales. “This requires a fundamentally different approach to security.”
AI agents can create a broader attack surface than conventional software because they may dynamically interact with multiple systems, data sources and other agents. Risks include prompt injection, leakage of sensitive information, unsafe outputs and actions taken outside an agent’s authorised remit.
Thales said its AI Security Fabric is designed to apply controls around agents, limiting access to data and blocking actions that fall outside predefined policies.
One example cited by the company is an AI agent used to settle insurance claims. An agent could potentially draw on personal information from an unauthorised source when determining a payout, creating privacy and compliance risks.
The security platform is intended to provide visibility into such interactions and allow organisations to enforce policies and detect AI-specific threats in real time.
“As agents become more capable and connected to critical data and systems, customers need security and governance built into how they operate,” said Vineet Bhan, Director of Security and Identity Partnerships at Google Cloud.
The companies said the expanded collaboration will also support organisations in assessing AI-related risks, protecting sensitive information and monitoring whether AI activity remains aligned with organisational policies and user intent.
There’s plenty of other editorial on our sister site, Electronic Specifier! Or you can always join in the conversation by commenting below or visiting our LinkedIn page.
