Blog
What your AI assistants can actually reach.
Plain writing about the AI assistants people install, what they can reach once installed, and what actually stops them.
-
Governance ·
Your AI rollout started before anyone approved it
Before writing an AI policy, establish what the agents on your company’s Macs can already reach.
Read · 1 min
-
Governance ·
The exception process becomes your security policy
Who can release a blocked action, for how long, and why determines whether security controls survive everyday work.
Read · 2 min
-
Operations ·
A quiet dashboard does not prove your agents are safe
Silence means little until you know the controls are running. Verify refusals, legitimate allowances, and reporting freshness.
Read · 2 min
-
Incidents ·
Nine seconds
A coding agent deleted a production database and its backups. There was a rule, in capital letters, and the agent had read it — it just named a different kind of destructive command. That is the part worth sitting with.
Read · 7 min
