Transparence sur l’IA

Where AI is used

Lifentry may use AI in a limited way: an optional assistant that helps with your tasks, and extraction of fields from documents. AI features are off unless the owner switches them on and, at the time of this version, are not enabled for public use. Text in digital PDFs is read locally without AI; scanned documents and photos are not sent to an external OCR service, so you enter their details manually.

A screen that lets you talk to an AI system says so before the conversation starts. An AI assistant is not a human adviser, a financial or tax adviser, a lawyer or a public authority.

Limits and human oversight

AI can be wrong or out of date. It works only with fields you confirmed, and raw document text is not sent to the assistant. Any action it proposes is limited to an allowlist of intents, is shown to you as a dry run first, and takes effect only after you approve it; you can revoke it. Contract explanations are not legal advice.

Verify anything important before relying on it for a payment, a deadline or a filing. The plannable amount is a deterministic calculation, not an AI output.

Your data and AI

Your content is used to deliver the feature you asked for and is never used for model training. Requests go through a gateway with a scoped key, a per-request budget and a prompt registry; we do not give model providers permission to use your content to train their own models. You can view and delete saved AI context. Model providers, processing regions and retention will be stated in the Privacy Notice before the feature is enabled. No solely automated decision with legal or similarly significant effects is taken about you, and a generic checkbox could not authorise one.

Reporting a problem

If an answer looks wrong or harmful, tell us at support@lifentry.net; a human reviews it, we correct errors and we may switch the feature off. You can ask for human review of any result that affects you.