For Microsoft security leaders, architects, and AI owners

Understand how Copilot works.
See how AI agents create risk.

"Security Is What Makes AI Work"

Hi folks, I’m Johnny.

I focus on securing identity, access, and data in the context of Copilot and AI agents, where risk isn’t always obvious, but the impact is real.

Most AI risk doesn’t come from attackers. It comes from perfectly valid access being used in ways organisations didn’t anticipate.

This site breaks down Copilot and Microsoft 365 architecture in a way that’s simple to understand. It also includes an agent simulator to show how weak security controls create real risk, along with writing on where AI and cybersecurity are heading.

AREAS OF FOCUS

Identity & Access Data Security (Purview) AI & Copilot Architecture Agentic Security

Keep scrolling to see what the website is about

Johnny Sheehan

Copilot Explained

Understand how Copilot really works.

Follow what happens when a user prompts Copilot, from identity and permissions through to data access, grounding, and response generation. See where security and governance controls apply along the way.

Agent Simulator

See how security decisions shape agent risk.

Explore how different security choices affect an agent across the build lifecycle. Test how permissions, approvals, access, and governance decisions can reduce or introduce risk.

Content & Insights

Explore articles, insights, and ongoing content.

Browse articles, breakdowns, and links to my content and socials, focused on the questions teams keep running into as AI and cybersecurity start to overlap.

Understand how Copilot works.See how AI agents create risk.

Understand how Copilot really works.

See how security decisions shape agent risk.

Explore articles, insights, and ongoing content.

Understand how Copilot works.
See how AI agents create risk.