Who approved that?
We bet you your AI agents have more access than you!
In 60 minutes Revolgy and JumpCloud show how to put AI agents under the same rules as your staff: who they are, what they can touch, and who signed off. Demo included.
Watch the recording
JumpCloud research · 800 IT leaders
give AI agents the same access as employees, or more.
have those agents written into an actual identity policy.
hour to see how to close that gap, demo included.
Sound familiar?
Nobody hired these agents. They still have keys.
Tools nobody registered
People connect AI clients to company data on their own. You find out later, if at all.
Keys everywhere
Shared API keys and service accounts sit in scripts and chats. No owner, no expiry.
Access that never ends
An agent gets access for one project and keeps it. Forever.
No name on the action
Something changed in a system. A person, or an agent acting for them? Nobody can say.
Leavers who don’t leave
The person who built it left. The agent still runs, still has access.
Who is accountable?
If an agent does damage, who answers for it?
The recording
What’s in the recording
The demo
Six steps, on screen, no slides
A JumpCloud engineer takes an AI client from "who is this?" to "here is exactly what it did", on screen in a real tenant.
The controls sit at the MCP layer, so the same approach applies across AI clients, not just agents on Google Cloud.
- 1 Admin registers an enterprise MCP server in JumpCloud AI Gateway.
- 2 Access is scoped to the right people and groups.
- 3 A user connects their AI client through one controlled endpoint.
- 4 The user authorises specific tools, and only those.
- 5 The AI client does real work in a connected business app.
- 6 The full trail shows up in Activity Log and Directory Insights.
Your hosts
Two teams who live in this stuff
Ashley Saunders
Hosts the session. Works with IT and security teams across the UK and EMEA on Google Cloud and Google Workspace.
LinkedIn →
Duncan Jaques
Bridges identity and access management with endpoint management. Specializes in automating user lifecycle, enforcing Zero Trust, and device-aware access across multi-OS hybrid environments. Runs the demo.
LinkedIn →The details
Questions? Write to marketing@revolgy.com
What you get
- The full recording, straight to your inbox
- The slides, to keep and share with your team
- A free Agentic AI Security Assessment, if you want one