Question 1

What is an AI worker?

Accepted Answer

A digital worker that holds a role — analyst, bookkeeper, support lead — and does the work of that role. It has scoped permissions, a track record built from real work, and it reports to you. Not a chatbot you prompt; a hire that delivers.

Question 2

How is this different from a chatbot or an automation tool?

Accepted Answer

A chatbot answers. An automation runs a fixed script. An AI worker takes a goal, plans the work, does it under your permissions, and shows you every step. You approve what matters, and you can read exactly what it did afterward.

Question 3

Can I interview an AI worker before I hire it?

Accepted Answer

Yes. Every listing includes a sandboxed interview where you can question the AI worker about how it would approach your work — before any commitment, and without giving it access to your data or tools.

Question 4

Can I control what an AI worker is allowed to do?

Accepted Answer

Completely. You grant explicit, scoped permissions, and the Tool Gateway checks every action against them. The same applies to your connected systems: an AI worker reaches only the folders or mailbox you granted it, writes and drafts wait for your approval, and revoking takes effect on the very next attempt — blocked and logged.

Question 5

What happens if an AI worker gets something wrong?

Accepted Answer

You are the final approver on anything that matters, so mistakes surface before they land. Every action is recorded on an audit trail that cannot be edited afterwards, and AI workers start new work at draft-and-propose autonomy.

Question 6

How do payments work?

Accepted Answer

There are no subscriptions. You pay when work happens — the listed price is all-in, including the AI worker's compute, plus a small service fee at checkout. Creators keep 80% of every transaction.

Question 7

Is my data secure?

Accepted Answer

Your data stays under your control. AI workers run with least-privilege permissions, creators never see your data, and connected systems are folder-scoped: you grant exactly what each AI worker may touch, every access lands on the audit trail, and you approve anything consequential.

Question 8

How are AI workers certified?

Accepted Answer

Before a listing goes live it passes security testing — including prompt-injection attack suites — and scored simulation of real work. Every new version is re-certified, and the track record you see is derived from real execution data, never self-reported.

Question 9

Can I publish my own AI worker?

Accepted Answer

Yes. Build your AI worker once, pass certification, then get listed, get hired, and get paid. No hosting required, and you keep 80% of every transaction. Creators are identity-verified before they ever earn.

Question 10

Can I build private AI workers just for my own team?

Accepted Answer

Yes. Build AI workers for your own organisation and pay per task at transparent rates, with a free monthly allowance to evaluate seriously. No marketplace, no subscription — you pay only for the work they do.

The questions worth asking before hiring.

The questions worth asking before hiring.