Hazonlive
DeutschEnglish
Sign in

AI transparency notice

Hazonlive uses artificial intelligence to support you in your day-to-day work. On this page we explain transparently how we use AI.

Which AI systems do we use?

Hazonlive uses large language models (LLMs) and image-generation models from several AI providers to run the assistants:

  • Anthropic Claude — Text AI, AI agents, document analysis, conversation processing.
  • OpenAI (text) — Text AI as a fallback provider when Claude is briefly unavailable, and for assistants for which OpenAI is selected in your workspace; also semantic search in your knowledge base.
  • OpenAI (images) — AI image generation on user request. The image description is transmitted; if you have an image revised, that image as well, and if an image is to show a person from your media library, the reference photos you marked for this.
  • Google Gemini — Backup image model, used only if image generation at OpenAI fails. The image description and up to three images recently AI-generated in your workspace are transmitted as a style reference.

These models generate text and images based on your inputs and the company knowledge you have stored. Which data goes to which provider is documented per processor, including purpose and data flow, in our privacy notice.

What is AI-generated?

  • All responses from the assistants (Anna, Markus, Julia, Lena) are AI-generated.
  • Email drafts, marketing copy, sales copy and service replies are created by AI.
  • Summaries, analyses and suggestions are AI-based.

Inside Hazonlive, every piece of AI-generated content is marked with this transparency badge (Hazonlive adds no AI notice to published posts):AI-generated

How do we ensure quality?

  • Human in the loop: By default, AI drafts wait for your review and approval. Social media posts that Lena schedules for a time (she is instructed to do so only when you give her one in the chat) and posts from Autopilot (off by default) are published at the scheduled time without a separate approval.
  • Guidelines: You can define rules per assistant that steer the AI's behaviour (e.g. tone of voice, prohibited topics).
  • Company knowledge: The AI draws on the company knowledge you store in the knowledge base. For research – such as trends or companies – an assistant also retrieves public web sources.

Data protection

  • Your data is stored in the EU (Frankfurt).
  • Your inputs are not used to train AI models.
  • Processing is carried out in compliance with the GDPR.
  • You can export or have your data deleted at any time.

Limitations

AI systems can make mistakes. Generated content may be inaccurate, incomplete or inappropriate. Therefore:

  • Review all content before using it.
  • Do not use AI-generated content as a substitute for professional advice (legal, medical, tax).
  • Responsibility for the use of the content lies with the user.

Contact

Questions about how Hazonlive uses AI? Write to us: kontakt@hazonlive.com

This notice describes where Hazonlive uses AI and how AI-generated content is labelled inside the app.