Security and control
What is proven in the AdoloGTM code today. Anything not yet decided is stated as such.
How we work
People decide
The AI agent proposes actions. No message is sent on your behalf without approval.
Agent controls
The agent controls menu has a switch per function, a Propose or Automatic choice, an emergency stop, and a daily budget ceiling.
Work screens sit behind sign-in
The dashboard and all data open only after signing in. The session is kept in a signed httpOnly cookie and expires.
Read-only and client accounts
Monitor, AE and client accounts cannot change data. A client account sees only its own data.
Demo mode writes nothing
Demo mode uses labelled sample data and does not write to the database.
Missing numbers are shown as missing
Data that is not yet measured reads "not yet measured", never made up or shown as zero. A closing counts only with proof of payment.
AI model
Claude by Anthropic for the AE functions.
Certifications, encryption at rest, and data retention period
Not yet decided.
Found a security issue? Write to the team on the Adolo WhatsApp (button below). A person on the team replies.
Talk to the team on WhatsApp