AI security
AI security, from first steps to a full program
Whether you're turning on ChatGPT for the first time or building AI into your product, start here. Pick the path that fits you.
Just getting started
For anyone: owners, managers and teams turning on AI tools for the first time.
I run IT
For whoever sets up and runs the AI tools.
- Security settings guides by productThe top 10 settings for ChatGPT, Claude, Gemini, Microsoft 365 Copilot, GitHub Copilot, Cursor.
- AI security hardening checklist21 steps to prepare identity, data and agents for AI.
- Free AI Hardening CheckChecks your AI admin settings and the agents on your machine, inside your AI assistant.
I'm the CISO
For security leaders building an AI security program.
We build with AI
For CTOs, developers and AI startups.
- AI security for CTOs: 50 questionsBAAs, PHI and PII, gateways, coding agents, MCP and incidents.
- Agents you buildHow to secure the agents and AI features in your own product.
- Security for AI startupsEnterprise readiness, model security and the EU AI Act.
- SOC 2 for AI startupsTraining data, model and API controls for SOC 2.
Turning on AI for your team:
six decisions
Your people will use AI either way. Give them a safe, fast way to do it, sized to the risk your data carries.
Step one
Pick your setting
Company accounts on a short approved list, a one page data policy, and new tools approved quickly.
Security settings guides
The 10 settings that matter most for each product, with where to find them and what good looks like.
Want the full picture?
The AI Security Assessment finds every AI tool, agent and key in use, including the AI apps with access to your email, files and code, and gives you a plan to fix what matters first.
See the AI Security Assessment