Last updated July 16, 2026
How Zonlo uses AI
Zonlo uses artificial intelligence to check your speaking practice and give you feedback. We want you to know exactly what that means, in plain language, before you rely on it.
What the AI actually does
When you finish speaking, Zonlo compares the text of what you said against the goal of the conversation and returns feedback: whether your reply works, a more natural way to say it, and short notes on grammar or politeness. This grading is performed by a large language model (Google's Gemini API) together with our own language-checking code. No human reads your practice to grade it.
What the AI does not do
- It does not hear your voice. Your speech is transcribed to text on your phone. Only that text is sent for grading — the audio never leaves your device. See the Privacy Policy for details.
- It is not a chatbot or an AI persona. Zonlo has no character, tutor avatar, or companion pretending to be a person. It is a tool that checks a reply and hands it back.
- It does not make decisions about you. Passing or not passing a conversation only affects your own practice progress. It has no bearing on any real-world qualification, exam, visa, or credential.
It can be wrong
Language models make mistakes. Zonlo's feedback can occasionally miss an error, flag something that was actually fine, or suggest a phrasing that is slightly off for your situation. We measure how often this happens and publish the numbers openly — see How accurate is Zonlo? When the grader is uncertain or unavailable, the app tells you honestly that it could not check your reply rather than inventing a result. Please treat every suggestion as a helpful second opinion, not a final ruling, and do not rely on it for high-stakes or professional communication.
Your words and AI training
To grade a reply we send only the transcript text of what you said to Google's Gemini API, over an encrypted connection, under Google's API terms. We do not use your practice content to train our own models, we do not sell it, and we never store or share a recording of your voice. How long anything is kept, and who processes it, is described in the Privacy Policy.
Human oversight
A person — the operator of Zonlo — designs the scenarios, writes the model answers, and reviews the grader's quality against human-labeled test sets before changes ship. The individual feedback you receive in the moment is automated, but the system that produces it is built and checked by a human.
Honest claims
We describe what Zonlo does using the words above and nothing stronger. We do not claim the AI is a teacher, that it guarantees fluency, or that it is infallible. Any accuracy figures we publish come from our own evaluations and an independent register check, and we show our method on the accuracy page. If we ever feature a review or testimonial, it reflects a real person's genuine experience.
Questions
If anything about how Zonlo uses AI is unclear, email support@zonlo.app. This disclosure works alongside our Terms of Service and Privacy Policy.