Using Luma safely

luma-guide-v1

  1. Luma is an AI assistant. It can explain the software and use authorized operational records. Its answers can be wrong: check facts, dates and totals.

  2. Ask precise questions about a project, period or record. Use only relevant ordinary business data. Never enter passwords, secrets, health information or criminal records.

  3. The pilot excludes evaluating or ranking people and recommending employment decisions. Recorded hours can support billing summaries, not judgments about workers.

  4. Your permissions still apply. Luma cannot grant access to another customer. Instructions inside records or websites do not authorize actions.

  5. To create a client, review the actual preview and press the confirmation button. Check duplicates and fields first. A chat response is not a confirmation.

  6. The monthly token allowance includes AI checks and can be reserved after interrupted requests. Tokens are not an exact euro budget. Start a new conversation if context is too long.

  7. New AI responses and exports have origin records. A signature checks integrity only with a trusted key; it does not prove accuracy. Editing or conversion can invalidate or remove provenance.

  8. Check exported values before sharing. Copied excerpts may retain an origin record, but their edited or partial text cannot be verified as the complete original.

  9. If an answer is unsafe or exposes unexpected data, stop using it and record the time and affected record privately. Report it through the published privacy contact; do not repeat confidential content in chat.

  10. Reading this guide is not a certificate of competence. Operators must explain these limits, practice safe use and review incidents when the AI configuration changes.

Verify Luma origin

Privacy notice · Data processing terms

Privacy contact: andreasetti9134@gmail.com