Fortinet adds AI security to its portfolio with the FortiGuard Inline AI Security Service. FortiGate customers get cloud-based inspection and guardrails for their AI traffic.
Fortinet is introducing the FortiGuard Inline AI Security Service. It is a cloud service that brings AI-specific inspection and guardrails to FortiGate firewalls. The service builds on FortiAIGate, which Fortinet is expanding with technology from the recently acquired Virtue AI.
Traditional security tools look at IP addresses, signatures and payloads, but do not assess prompts and responses. According to Fortinet, threats to AI sit in prompts, model responses and communication with MCP tools. They can indeed trigger harmful tool and API calls or actions by AI agents.
Real-time AI protection
The guardrails check prompts and data in real time during interactions with AI agents, large language models and MCP tools. The inspection covers text, images, audio, video and code. According to Fortinet, data loss prevention recognises 35 types of personal data and confidential data. Based on such detections, organisations can warn, block or mask data.
Administrators set model permissions, budgets and limits per organisation, workspace, project or user. Once a budget runs out, the system blocks requests. The security system stops so-called Denial-of-Wallet attacks before they consume tokens or GPU compute. Smart routing sends each request to the model that fits the content, policy and cost.
Teams get their own workspaces with their own policies, logs and permissions, under central management. More than 50 predefined policies on AI governance, privacy legislation and sector rules can be activated quickly. Auditors get logs of every prompt, output and tool call. The service also shows which AI agents are active and how they work with applications, APIs and MCP servers.
No separate infrastructure
According to Fortinet, organisations no longer need to buy extra AI security. The service runs via the FortiGate infrastructure that customers already use, so they do not have to roll out or manage a separate environment.
Customers with stricter privacy and sovereignty requirements can host FortiAIGate themselves, on-premises or in their own cloud, on Kubernetes with Nvidia GPUs. Prompts, responses, data and logs then stay within their environment. Fortinet has not yet shared anything about pricing and availability.
The solution responds to growing demand for AI security. Security specialists are keen to play a role in it, even though protecting against AI misuse is a different trade from traditional cybersecurity.