How we handle access & data

Agents and automations get access only to the specific systems their job requires — never blanket admin rights. Where a tool supports scoped API keys or restricted roles, we use them.

Credentials are exchanged through your password manager or a secure vault, never over email or chat. Access is documented, and you can revoke it at any time without breaking your own systems.

Anything customer-facing, financial, or contractual passes a human approval checkpoint before it goes out. AI drafts and prepares; your team decides what leaves the building.

Approval boundaries are agreed in writing during scoping — you always know which actions run automatically and which wait for sign-off.

Automations keep a readable log of what they did, when, and why. If something looks wrong, you can trace it — and so can we.

Logs live in your accounts wherever possible, so the audit trail belongs to you, not to us.

We use business-grade AI services under agreements that do not permit training on your data. Your data stays in your accounts; models process it to do the job and nothing more.

Which AI provider processes what is documented per automation, so you can answer your own customers' questions about data handling.

Every engagement ends with documentation and handover: what runs where, with which access, and how to change it. If we part ways, our access is removed and your systems keep working.

No lock-in by design — automations are built on your accounts and tools, not trapped inside ours.