AI transparency disclosures
Last updated: 3 June 2026
This page describes the AI systems used in the BookOS platform and the transparency obligations BookOS observes under the EU AI Act (Regulation (EU) 2024/1689, “AI Act”). It supplements our Terms of Service § 18 and our Privacy Policy.
1. Our role under the AI Act
Where BookOS embeds an AI system in the Service, BookOS is the deployer of that AI system within the meaning of Art. 3(4) AI Act. Where you, as our customer, configure the AI feature (e.g. by writing custom prompts or training data into the Service), you may also act as a deployer in respect of your own use. We do not develop or place on the market any general-purpose AI model.
2. AI features in the BookOS Service
The table below lists AI-assisted features currently available in BookOS, the third-party model provider, the purpose, and the AI Act risk category (per Title III). All listed features are limited-risk or minimal-risk; BookOS does not deploy any AI system that falls within the high-risk categories of Annex III.
| Feature | Purpose | Model provider | Risk class |
|---|---|---|---|
| No customer-facing AI features are active at this time. Any future feature will be listed here at least 30 days before launch, together with its model provider, purpose, and applicable transparency notices under Art. 50 AI Act. | |||
3. Art. 50 transparency obligations
- Interaction with an AI system (Art. 50(1)): where an end user interacts directly with an AI-powered chat, copilot, or voice agent, we will clearly inform the user that they are interacting with an AI system, unless the AI use is obvious from the context.
- Synthetic content (Art. 50(2)): where the Service generates or materially modifies image, audio, video, or text that could be mistaken for authentic human-produced content, the output will be marked as machine-generated using a machine-readable watermark or label, to the extent technically feasible.
- Emotion-recognition / biometric categorisation (Art. 50(3)): not used. Should we ever introduce such a feature, we will inform affected persons in advance and process any personal data in accordance with applicable law.
- Deep-fake disclosures (Art. 50(4)): the Service is not used to generate deep fakes of identifiable persons. Customer use of the Service for that purpose is prohibited under our Acceptable Use Policy.
4. Training data and model improvement
BookOS does not use Customer Data, end-customer personal data, salon photographs, booking history, or message content to train or fine-tune general-purpose AI models, whether our own or those of third-party providers. Where a third-party model provider is used at inference time (e.g. through an API), we contractually require the provider not to use customer prompts or outputs for model training.
5. Human oversight
AI-generated suggestions are presented as suggestions, not decisions. A human user must review and accept an AI-generated message, schedule, or recommendation before it has any effect on a booking, payment, or customer-facing communication. Automated decisions that produce legal or similarly significant effects are not made — see Section 4c of our Privacy Policy.
6. Reporting AI-related concerns
To report a suspected bias, factual error, harmful output, or other AI-related concern, email admin+ai-concerns@bookos.dk. We acknowledge receipt within 5 business days and investigate diligently. Reports of incidents involving serious harm to fundamental rights will be escalated to the relevant market-surveillance authority where required by Art. 73 AI Act.
7. Updates
This page is reviewed each time a new AI feature is added to the Service and at least annually. Material updates will be announced to billing contacts at least 30 days before they take effect.