AI disclosure
Last updated 1 November 2026. Where we use AI, what our agents do on their own, and what they never do.
Our agents are AI
Every agent in this store uses an AI model for some of its work: reading, sorting, scoring, summarising or drafting. Each agent's page says which parts are AI and which are plain rules.
You approve external actions
By default, anything that leaves your business, such as an email or message to a customer, a payment, an offer, or a public post, waits for a person to approve it. You can loosen this for low-risk steps, and the agent's setup guide shows how and what the risks are.
AI can be wrong
AI models make mistakes. Our agents are built to say when they aren't sure, to fall back to simple rules when the AI is unavailable, and to log what they did so you can check it. Please don't use them as the only safeguard for decisions that matter.
Rules we build in
- Voice agents tell callers up front that they're AI and that the call is recorded. See AI and call recording.
- We don't build agents that make automated outbound sales calls.
- No scraping of personal data or of anything behind a login.
- Hiring agents only draft and organise. A person makes every hiring decision.
- We follow the usage policies of the AI providers we build on, and we don't build for healthcare.
- You own your accounts. We never ask for your passwords.
How this site was made
AI tools helped write the code, the agents and the words on this site. A person reviews and approves what ships. Demos use fictional businesses and data, and say so.