Short answer
A FirstNet Secure AI project has four steps: discover, design, deploy and manage. Most start with an AI security assessment for one team.
In detail: How a Secure AI with Cloudflare Project Works
The four steps:
- Discover: AI usage discovery and a risk assessment, covering which AI tools, agents and data are in play and where the gaps are.
- Design: policy and guardrail design, including approved tools, data rules and access per team, agreed with your risk owners.
- Deploy: integration with your identity provider, network, SIEM and applications. Controls start in logging mode and move to blocking once the rules are proven, so nobody is cut off by surprise.
- Manage: ongoing monitoring and a managed service, with policy changes, reporting and escalation into Cloudflare.
From there you can add prompt and data controls, AI Gateway for your own applications or MCP server portals for AI agents as the need grows. A FirstNet specialist can scope the assessment with you.