AI transparency
This page explains how the public support chat and the protected customer project chat use artificial intelligence, translation and human approval.
The support chat helps you understand the two primary service areas (DATEV/e-invoicing automation and NIS2/BSIG technical data resilience) plus supporting modules. It explains technical scope, boundaries and typical next steps and can collect minimised context for a request.
In the customer portal, a customer can write in their own language. The server keeps the original, prepares a Russian translation for the administrator and creates concise Russian reply drafts using the relevant service, order, source page and processing stage. The administrator chooses, edits or writes the final Russian answer; only then is it translated back into the customer language and released.
The AI does not provide legal, tax or individual professional advice. It does not replace a contract, invoice check or technical audit. For binding decisions, deadlines or compliance questions, please speak with a qualified person.
AI reply suggestions are drafts and are never sent automatically. An administrator remains responsible for the final project-chat answer and can ignore, edit or replace every suggestion. Other automated service outputs also start in draft or pilot mode and retain their documented approval gates.
Project-chat messages, translations, selected context and delivery status are stored in the protected portal so both parties can follow the work. Conversations with no new activity are removed regardless of status after the configured retention period (default: 365 days); customer-account deletion also removes its chat threads. OpenAI API requests are made only by the server with API conversation storage disabled; provider security/retention terms and our privacy notice still apply. Do not share passwords, API keys, full payment data or unnecessary personal information in either chat.
Questions about this AI disclosure or a human follow-up: