Trust
The AI cannot see more than you can.
This is the single most important sentence on this site about the AI, and it is a structural property rather than a policy: the assistant queries as the person asking, through the same permission checks.
Inherited, not configured
There is no separate AI permission set to keep in step with the real one. The question runs with the asker's identity, so the two cannot drift apart.
Refusals say what happened
Asked something they may not see, a user is told the information exists and they do not have access — rather than being told it does not exist, which teaches people the system is unreliable.
Content is data, never instruction
Text inside a document, a form submission or an imported record cannot change how the assistant behaves. Anything a person typed is treated as information about the business, not as a command.
Trust Centre
The rest of this area.
Security, employee privacy, AI permissions, audit and availability — stated plainly.
Security
Encryption, tenant isolation, production access and credential handling.
Employee privacy
What the people in the system can see, and what is kept from whom.
Audit trail
Who changed what, who approved it, and who looked at the sensitive things.
Availability & continuity
Backups, recovery and what happens on a bad day.
Data processing
The roles, the instructions and the subprocessors.
Run your team with clarity.
From the moment someone joins to the day their responsibilities are handed over, Orglume keeps your people, their work and your organization connected.
Set up from a spreadsheet · Nothing about the AI is metered · [email protected]