Platform
Built to be supervised first, autonomous later
Every Aiconomy agent runs on the same platform: the same approval controls, the same integration layer and the same escalation logic.
Four steps to go-live
1. Connect your tools
Gmail, WhatsApp, Xero, your calendar and whatever else you already use, via secure OAuth.
2. Teach the basics
Tone of voice, key policies and a short list of things the agent should always check with you first.
3. Run in approval mode
The agent drafts every action. You approve, edit or decline; nothing goes out unsupervised.
4. Switch to autopilot
Once you trust the pattern, flip specific actions to run automatically. You can revert any time.
Human-in-the-loop, by default
Nothing sends, posts or pays without a human decision until you say otherwise. Every action is either informational (reading, summarising, flagging) or consequential (sending a reply, posting content, releasing a payment), and consequential actions are gated behind approval.
Weeks 1–2
Every consequential action is drafted and held for your review before it happens.
Week 3 onward
You selectively approve specific action types to run automatically, one at a time.
Any time
Revert any action type back to approval mode instantly.
Security and compliance
- UK/EU-hosted infrastructure where possible: data is encrypted in transit and at rest.
- ICO registered: Aiconomy is registered with the UK Information Commissioner's Office.
- EU AI Act aware: customer-facing agents disclose they are AI. Read the regional compliance position and our Article 50 guide.
- No training on your data: your business data is never used to train underlying AI models.
Questions
Does anything send, post or pay without my approval?
Not by default. Every action is classified as informational (reading, summarising, flagging) or consequential (sending a reply, posting content, releasing a payment). Consequential actions are held for your approval until you choose otherwise.
Is my data used to train AI models?
No. Your business data is never used to train underlying AI models.
Do agents tell people they are AI?
Yes. Customer-facing agents disclose that they are AI at the start of a conversation, in line with current transparency requirements. See our guide to EU AI Act Article 50.
See it applied
Every agent page covers what the agent does, what it connects to and what it costs. Meet the team.
Try an AI employee for 14 days
£99 for 14 days with one agent of your choice. If you subscribe within 7 days of the trial ending, the £99 is deducted from your first invoice.