On a white-label account your clients see your name, not ours. That includes the AI features: the AI agents that draft and post replies to reviews, and translation, all run under your brand. So your own privacy policy is the one that has to tell your clients about them, not ours.
An agent's Behavior section showing the mode and the rating bands a client can read.
This page gives you wording to adapt. It is a starting point, not legal advice: have your own counsel review whatever you publish, the same way we review ours.
What to cover
Your privacy policy should tell clients that:
- Your product includes AI features they can switch on, such as reply drafts written by an AI agent, and translation of reviews and replies.
- An agent writes from the review, the client's own instructions and the knowledge they add. In draft mode nothing is posted without a person; in auto-reply mode it posts only at the star ratings the client chose.
- Private feedback is never posted by an agent. It drafts a reply for a person to send by email.
- Only text is involved: the review, its title and rating, the reviewer's name and the knowledge passages that matched.
- The AI processors behind these features. We keep the current list, by category, with what they receive and where they run, on an unbranded page at reviewintegration.com/privacy, and we name each AI subprocessor on request. Link to it instead of naming providers yourself: it changes only there, and your policy stays correct without an edit every time it does.
Template wording
Copy this into your own privacy policy and adapt the bracketed parts:
AI features. [Your agency]'s product includes AI features you can switch on, including reply drafts written by an AI agent and translation. An agent writes from the review, your instructions and the knowledge you add. In draft mode nothing is posted until a person approves it, and in auto-reply mode replies are posted only at the star ratings you choose. Private feedback is never posted. Only text is involved, never images, video or attachments.
AI processors. These features run on a small set of AI providers under contract, who process your content only to return a result and do not use it to train their models. The current list, by category, with what they receive and where they run, is published on our processors list, and each AI subprocessor is named on request.
If our privacy page changes what it says
The "no training" line above matches what our own privacy page says today. If that page ever changes it, update your own wording to match, since a client reading your policy is relying on it being accurate.
What this costs
Nothing. Publishing your own policy is up to you, and the AI features are priced in credits the same way whoever's name is on the product.