What people ask before they start

In an EU region (Frankfurt), every company on its own, separated rows; files in private storage. The AI model (Claude by Anthropic) does not train on the data. Details in the privacy notice and on the security page.

No. The agent has no send, delete or pay tool. It writes a draft; the approval is bound to the exact content that goes out, and if anything changes it asks again.

Monthly periods, cancel any time for the end of the period. Then everything stays readable and exportable for 60 days, then we delete it. No annual commitment, no setup fee.

No card. For 14 days you get the full Átadás plan with €5 of AI. At the end you pick a plan, or the workspace stays read-only.

Because Stafeta holds the company's mail and customers. A password (or Google/Apple) and an authenticator app code – for the shift worker too. A minute to set up, and most competitors cannot enforce it.

On your own: the industry pack gives you a working board on the first morning and the tour walks you through. If you want us to put your tasks in lanes and train the team, there is the Warm-up, the Scan and the System.

E-mail through a forwarding address; replies from your own Gmail or Microsoft account with send-only rights; Mailchimp campaigns from the card; Drive and Docs read and create; WhatsApp and voice messages from the phone. A new connector only comes in with narrow rights.

The platform produces the AI Act Article 4 record itself and we give the team training material; for the GDPR there is a data processing agreement and a sub-processor list. Our legal pages are open and a lawyer reviews them before launch.

A company price in three sizes, by the number of active people. You move up only after two consecutive months over the limit, down any time. Agents do not count as people.