AI review response generator

How the generator works

Paste the text of a review (up to 1,500 characters), then set the star rating, your type of business and the tone you want. The tool sends only that information to a language model with strict instructions, and shows two reply options worded differently, each with a copy button.

The AI follows the same rules as Revys AI Auto-Reply:

• Nothing made up. The reply only uses what the review says: no customer name, no staff names, no promises, discounts, phone numbers or email addresses. Any option that contains a link or contact detail is dropped before it's shown.
• The review is data, not instructions. If the pasted text contains commands (“ignore your rules”), the AI ignores them; text that clearly isn't a review is refused.
• Negative review (1 to 3 stars): empathy, an apology where it fits, and an invitation to continue the conversation privately. Never defensive, and never an admission of fault or liability.
• Positive review: a genuine thank-you that echoes what the customer liked, without filler.

Nothing is stored: not the review, not the reply, not your IP address in a database. A temporary in-memory counter limits the tool to 3 replies per day per connection, and 200 per day across all visitors.

How to reply to a negative review

A review reply is public: it speaks to future customers as much as to the reviewer. Google recommends replying in a courteous, helpful way. A few simple rules, which the tool applies and which you can check when you read the draft:

1. Thank them for the feedback, even if it feels unfair: hesitant customers read your reply.
2. Acknowledge the experience described without arguing the details in public.
3. Offer a concrete next step in private: a call or an email, using the contact details on your profile.
4. Keep it short: three to five sentences are enough. A long justification does more harm than the review itself.
5. Never post personal information about the customer, even if they identified themselves.

If a review breaks Google's rules (abuse, conflict of interest, fake review), replying isn't enough: report it from your Google Business Profile.

Sensitive reviews: health, legal, safety

Some reviews call for extra care. In professions bound by confidentiality (doctors, dentists, physiotherapists, pharmacists, lawyers…), writing “during your appointment on 12 March” already confirms in public that someone is your patient or client. The generator spots these cases (a healthcare or legal business, or words such as “allergy”, “lawyer”, “complaint”, “police”) and then writes a deliberately general reply: thank you for the feedback, we take it seriously, please contact us directly. It doesn't confirm any relationship and doesn't comment on the facts.

Even in careful mode, read the reply: only you know whether a detail could identify someone or commit you to something.

What the tool doesn't do

• It doesn't post anything or connect to your Google profile: the reply is a draft for you to check.
• It doesn't know your business: no name, opening hours or context. Add an accurate detail yourself if it helps.
• AI can misjudge the tone or miss sarcasm: a human read-through is still essential.
• It handles one review at a time, up to 3 per day.

To keep up with every review, Revys fetches your Google reviews each night, drafts a reply in your tone with your sign-off and your banned words, and lets you approve it before it goes live. Sensitive reviews are never posted without a human check. Your reviews can also be shown on your website in a format that Google and AI search can read.