1. Purpose and Scope

1.1 Document Purpose

This AI Transparency Notes document provides a factual disclosure of the FortiAI feature embedded within FortiManager. It describes how the AI assistant works, what data it processes, how it is designed, and what safeguards are in place - enabling administrators, compliance teams, and auditors to make informed decisions about its deployment and use.

All content in this document is based directly on the FortiAI FortiManager product specification.

1.2 What is FortiAI on FortiManager?

FortiManager delivers a single-pane-of-glass management experience spanning connectivity, resource utilization, device settings, policy status, and alerts. It orchestrates FortiGates across varied form factors - VM, CNF, standard and ruggedized hardware - and extends management to the broader FortiOS-based networking ecosystem including FortiSwitch, FortiAP, and FortiExtender, ensuring consistent enforcement of unified security policies across the organization.

While FortiManager is a powerful platform, it carries a learning curve and some operations can become complex. FortiAI on FortiManager is an intelligent assistant designed to reduce that complexity: to simplify operations and to bring features that might otherwise be multi-step, complicated, or impossible to accomplish via the regular UI, while keeping security as a top priority built in from day one.

FortiAI provides a chat window within the FortiManager UI through which administrators can ask questions and request tasks in natural language. Answers and results are returned in the same chat window. Specific features can also be triggered directly from the UI via dedicated buttons - for example, explaining the difference between two configuration revisions, or explaining a planned installation.

FortiAI specifically harnesses Generative AI to automate routine tasks, optimize configurations, accelerate diagnostics, and streamline policy management. It integrates with FortiAIOps to support predictive network analysis and automated troubleshooting, enabling administrators to proactively identify and resolve potential issues before they impact network performance.

1.3 In-Scope Features

Agent / Feature	Description
CLI and Jinja Script Agent	Turns plain-language intent into CLI or Jinja scripts for FortiGate and FortiManager. Supports creating, modifying, and explaining scripts - whether starting from scratch or refining existing ones - and guides toward best practices.
Device Management Agent	Provides real-time device health, licensing, firmware status, upgrade readiness, and security risk assessment. Explains configuration changes in plain language - including the CLI commands used and their intent. Supports searching devices by interface configuration. Gives visibility into recent changes, installation activity, and devices needing attention.
Firewall Policy Agent	Enables natural language search, inspection, creation, editing, cloning, version comparison, and rollback of firewall policies and related objects across the FortiManager environment.
VPN and SD-WAN Network Diagnostic Agent	Diagnoses and resolves VPN, SD-WAN, performance, and device issues through natural conversation. Analyses symptoms, walks through root cause step by step, and guides administrators to restore network stability - without requiring deep technical knowledge upfront.
Provisioning Template Agent	Guides SD-WAN overlay design, site connectivity, and traffic flow optimization. Creates and manages reusable templates for BGP, IPsec VPN, and system settings. Highlights configuration differences and conflicts before changes are applied. Supports device onboarding, variable management, and consistent deployment at scale.
General Knowledge Agent	Answers product questions, compares features and specifications, navigates to specific UI pages within FortiManager, and locates relevant Fortinet documentation. Serves as a guided entry point for FortiManager and Fortinet product knowledge.

The agents can either work alone to solve problems in particular areas or team up to tackle more difficult ones.

1.4 Out-of-Scope

• Autonomous configuration changes without explicit administrator confirmation - all data mutations require user approval

• Processing of end-user personal data or enterprise business content unrelated to network management

• FortiAI does not modify user input or raw data beyond the masking and unmasking of sensitive fields described in the following topic: 4. Data Flows, Protection, and Retention

• MCP Server for public or third-party use - not yet released; Fortinet reserves the right to release it in a future version

1.5 Target Audience

Audience	Relevance
Network Security Administrators	Primary users; interact with FortiAI daily for device management, policy work, and diagnostics.
Security Operations Teams	Consumers of AI-generated diagnostics, reports, and remediation guidance.
IT Compliance and Audit Teams	Review of AI governance, data handling, and acceptable use boundaries.
CISOs / IT Leadership	Strategic oversight of AI deployment and associated risk posture.
Data Protection Officers	Privacy impact review; data flows, masking practices, and retention policies.