Powerful agents. Hard limits.
Every agent works inside rules you set. It reads what you allow, acts where you approve, and leaves a trail of everything it did.
- Every action logged
- Keys in a vault
- EU and GDPR
What an agent may do
| Read orders and tickets | On its own |
|---|---|
| Answer a customer question within policy | On its own |
| Update a CRM record | On its own, logged |
| Issue a refund above your limit | Only with your approval |
| Send a campaign to your whole list | Only with your approval |
| Delete customer data | Never, only a person |
Example. You set the limits per agent during the build.
Six rules every build follows
Your data
Your data, logs and prompts are yours. Growth Systems, Automation and AI Systems run in our environment; only Custom SaaS runs in your own accounts.
Approved sources only
Agents read only the systems and files you give them access to. Nothing more.
A person decides
Anything outside your rules goes to a person before it happens.
Every action logged
Each step an agent takes is recorded: what it did, when and why. You can always trace it back.
Keys in a vault
API keys and passwords live in a secure vault. Never in code, never in a chat.
EU and GDPR
Hosting and storage stay in the EU, and we sign a data processing agreement with every client.