Tenable expands AI security coverage to Google Gemini

Tenable has expanded the AI security features in its Tenable One Exposure Management Platform, adding support for Google Gemini and broadening discovery across AI developer tools.

The release extends coverage across major large language model platforms, including Anthropic Claude, OpenAI ChatGPT Enterprise, Microsoft Copilot and Google Gemini. It also adds visibility into Model Context Protocol deployments, AI-native integrated development environment tools and AI-enabled browser extensions.

Tenable is positioning the changes around what it describes as a widening AI exposure gap inside organisations as staff adopt both approved and unapproved AI services. Its platform is designed to discover AI use across endpoints, cloud environments and LLM applications, and to track the assets involved through its Exposure Graph data system.

The business cited data from more than 7,000 organisations that identified 457 million AI-related security issues over a 30-day period, equal to an average of 62,000 exposures per organisation. The figure is intended to illustrate the scale of the visibility problem for security teams as AI tools spread beyond formal procurement and governance processes.

Broader coverage

Under the latest update, Tenable One now monitors Google Gemini usage with controls including oversight of user interactions and prompt responses, policy enforcement, and detection of malicious activity or inappropriate use. Tenable says this expands visibility and governance across the main LLM platforms used in business settings.

Another part of the release focuses on AI development and integration tools that can sit outside conventional security oversight. Tenable now supports major Model Context Protocol deployments and AI-native IDEs, including Cursor, Windsurf and Trae, as software teams bring AI assistants into daily development work.

The update also doubles coverage of sanctioned and shadow AI. In practice, this means security teams can map a wider range of approved services and unauthorised tools that employees may be using without formal sign-off.

Remediation tools

Tenable has also added workflow functions intended to help organisations respond to identified risks. Users can now create remediation tickets directly in Jira and ServiceNow, while policy violations can trigger automated notifications by email, Slack or Microsoft Teams.

These additions reflect a wider push among cybersecurity vendors to connect detection more closely with operational response. Rather than isolating findings inside specialist dashboards, vendors increasingly route issues into the systems where IT, engineering and security teams already manage work.

Tenable distinguishes between two AI-related elements within the Tenable One platform. AI Exposure is aimed at helping organisations discover, assess and secure their use of AI, while Tenable Hexa AI is described as the system's agentic engine for automating security tasks and coordinating remediation activity.

The update comes as companies face growing pressure to understand where AI systems are being used, which data they access and how they may create new paths for cyber attacks. The rise of connected applications, identity systems, cloud infrastructure and AI tools has made those risks harder to trace with older security products built around narrower parts of the technology stack.

Eric Doerr, Chief Product Officer at Tenable, outlined the company's rationale for the expansion. "The massive volume of AI exposures confirms the operational reality that authorised and unauthorised AI is deployed faster than security teams can govern it," he said.

He added: "There's no denying that AI attack surfaces are making defenders' jobs even harder, and legacy or siloed cybersecurity tools simply don't cut it. With today's expansion to include Google Gemini, MCP and AI-native IDE deployments, Tenable is the only exposure management platform delivering unified AI visibility and governance across all major LLMs, software, and tools."

Tenable says it serves more than 40,000 customers globally. The latest changes are intended to give security teams a more complete view of where AI is being used, what risks it introduces and where action is needed.