Fortinet launches AI security service for FortiGate

Fortinet has introduced the FortiGuard Inline AI Security Service for FortiGate customers, adding AI-specific inspection and policy controls through a cloud-delivered model.

The offering is intended to help organisations monitor AI activity and manage threats and data exposure through existing FortiGate firewall deployments. It also expands the role of FortiAIGate, Fortinet's technology for inspecting prompts, model responses and related AI traffic.

The launch comes as businesses move AI tools from experimentation into operational systems, where large language models and autonomous agents may interact with sensitive data, internal applications and business processes. That shift has created demand for security tools that inspect not only network traffic and files, but also the content and context of AI interactions.

The service allows customers to add AI inspection without deploying separate infrastructure. For organisations with stricter privacy or sovereignty requirements, FortiAIGate can also be deployed within a customer's own environment.

Fortinet said the cloud-delivered service inspects prompts and data in real time, applying security, compliance and data protection policies across interactions involving agents, large language models and Model Context Protocol tools. It can also discover shadow AI use and apply data loss prevention policies to AI traffic.

Broader controls

Fortinet also outlined a wider governance model for AI use inside organisations. Administrators can set permissions for model use, budgets, spending limits and rate limits by organisation, project, workspace or user, with alerts as spending approaches a defined threshold and blocks once a limit is exceeded.

The system is also intended to block malicious prompts and so-called denial-of-wallet attacks before they consume tokens or GPU resources. Requests can also be directed to different AI models according to content, policy and cost.

For governance teams, the system includes separate workspaces with distinct policies, logs, roles and permissions, while preserving central visibility. Customers can also apply more than 50 predefined policies linked to AI governance frameworks, privacy rules, industry regulation and security frameworks.

Fortinet said its expanded data loss prevention coverage identifies 35 types of personally identifiable information and confidential data. Customers can configure the system to alert, deny or redact that information before it is exposed.

AI production risk

Fortinet positioned the product around the limits of traditional security tools in AI environments. Conventional systems inspect IP addresses, signatures and payloads, but do not assess prompts and responses for threats that can be hidden in exchanges between applications, models and AI tools.

That is becoming more relevant as AI agents gain the ability to call tools, use APIs and take actions with a degree of autonomy. In that setting, the security challenge extends beyond data leakage to include harmful instructions, risky tool use and uncontrolled spending.

Michael Xie, Founder, President and Chief Technology Officer at Fortinet, said: "Organisations need AI security that can scale with how quickly AI is being adopted across the enterprise. By extending advanced AI inspection and guardrails through FortiGate, we are making AI security available through the infrastructure customers already use. FortiAIGate provides the foundation for this broader strategy, giving customers flexibility in how they deploy and consume AI security based on their cost, performance, privacy and sovereignty requirements."

The launch also reflects Fortinet's effort to build AI security into its broader product portfolio rather than position it as a separate stand-alone platform. The service connects AI inspection and runtime controls with its existing products across network, application, cloud, data, identity and security operations.

Fortinet also pointed to technology from its recent acquisition of Virtue AI as part of the FortiAIGate development path. The acquisition appears to support the company's work on AI guardrails, visibility and control functions as enterprises put AI systems into production.

Will Townsend, Chief Analyst at LoneStar Advisory & Research, said: "FortiAIGate takes an elegant approach to a complex problem. Rather than asking customers to manage separate tools for AI security, governance, cost and agent visibility, Fortinet is bringing these capabilities together in a single gateway. The rapid integration of Virtue AI technology further strengthens that story by improving the guardrails and performance customers need as they move AI into production."