About the AI and its answers

Last updated: 3 November 2026

You are using an AI system. allqus uses artificial intelligence (shown as “nosh GT-06”) from OpenAI and Google (Gemini) to search the web, write answers, documents and code, and make images, videos and speech. This page explains how it works, what it can’t do, and how your data is protected, following the transparency rules of the EU AI Act and the guidance of data protection authorities on generative AI.

How an answer is made

The AI first works out what you want (for example a comparison, a calculation or step-by-step instructions) and picks a fitting format. Then it searches the web, reads the results and writes an answer with links to the sources, shown as small badges with the site’s name. For weather and flights, allqus also asks specialised services. Simple sums, greetings and translations are answered without a web search, and when no sources are found the AI answers from its general knowledge; such answers show no sources. With your own files, projects or connected apps, the relevant parts of them are included so the AI can answer from them. The AI Agent, Coder and AI writing in Docs work the same way and mark their results as made by AI.

Key figures and option lists

Some answers highlight a key fact or a key figure (for example a total price) in a box, and “which is best” questions get a ranked list of options. Figures are calculated by the AI from the sources, and the ranking is the AI’s own assessment of the sources: it is never paid for and is not a test or recommendation by Teknocare GmbH. Check figures and offers on the linked pages before you rely on them.

Writing help and messages

The grammar check in the question box and Write a message use AI to correct, rephrase, shorten or translate text you wrote. The text is sent to our AI provider only when a correction is made or you press a button, and it isn’t stored. The AI can change what you meant, get names or numbers wrong, or translate imperfectly, so read the result before you send it. Messages you send with WhatsApp are your own messages: allqus doesn’t add a label to them, so tell the recipient yourself if it matters to them that AI helped you write.

Images, videos and voice

Images and videos you ask for are made by AI (Google Gemini models) and are labelled “made by allqus AI”. They may carry an invisible watermark that identifies them as AI-generated. In voice mode your speech is turned into text and the answer is read aloud by an AI voice; it is a synthetic voice, not a person.

Limits

Answers can be wrong, incomplete or outdated, and sources can be misread. Check important facts in the linked sources.
Answers are not professional advice. For health, legal, tax or financial decisions, ask a professional.
Prices, timetables and availability (for example flights) can change at any moment; the provider’s website is binding.
The AI can reflect bias in its sources, and can be wrong about real people. Tell us when an answer is harmful, unfair or wrong about someone.
Code written by Coder can contain bugs or security problems. Review and test it before you use it.

Safeguards

Every answer shows its sources, so you can check it.
Web pages and files are treated as untrusted data: the AI is instructed not to follow instructions hidden in them.
The AI providers’ safety systems filter harmful content, and our acceptable use policy forbids misuse.
If our main AI is unavailable, a backup AI from our other provider answers, so you get an answer instead of an error.
Code from Coder runs in a separate, sandboxed page that can’t read your account or stored data.
No decision with legal or similar effect on you is made by AI.
People on our team review reports, and you can always reach a person through the support page.

AI-written documents and content

The AI Agent and the AI writing in Docs create texts, tables and presentations for you. Review them before you use or publish them; you are responsible for documents you share. Where the law requires it, say that published content was generated with AI. The AI never invents facts about you on purpose: where it doesn’t know something it leaves a placeholder such as [date].

Interface languages

Menus and pages in languages other than English are translated by AI. The translations can contain mistakes; the English version of the legal pages applies. Your own questions and content are never sent for this translation.

Your data

Questions and the content needed to answer them are sent to our AI providers, which process them on our behalf and don’t use them to train their models. We don’t train AI on your data either. Please don’t enter sensitive data unless it is needed. Details in the privacy policy.