AI prompt: reply to a positive Google review
A copy-and-paste AI prompt that turns a 4- or 5-star Google review into a short, personal public reply: thank the reviewer by first name, mention what they actually praised, pass thanks to the team member they named, with no incentives, keyword stuffing, invented details or promises.
When to use it
A happy customer has left a 4- or 5-star Google review and you want to reply quickly with something that sounds like a person wrote it for them, not a copy-pasted thank-you or an advert stuffed with your services and town.
- Copy the prompt below.
- Replace every
{{VARIABLE}}with your own details (see the list under the prompt). - Paste it into the AI chat tool you already use.
- Read the result against the checklist before you use it.
The prompt
You are helping a local business owner reply to a positive Google review. Write the public reply the owner will post under the review.
BUSINESS
- Business name: {{BUSINESS_NAME}}
- Type of business: {{BUSINESS_TYPE}}
- Who signs the reply (first name and role): {{SIGNER}}
- Tone: {{TONE}}
- Team members the reply may name: {{STAFF_OK}}
- Anything the reply must not mention: {{DO_NOT_MENTION}}
- Opening lines of our recent review replies (do not reuse these words): {{RECENT_OPENINGS}}
THE REVIEW
- Reviewer name as shown on Google: {{REVIEWER_NAME}}
- Star rating: {{STARS}}
- Review text: {{REVIEW_TEXT}}
RULES FOR THE REPLY
1. The first sentence starts with the reviewer's first name only (never their last name or initial) and thanks them for their kind words about one specific thing or person they praised. Do not thank them for "sharing" or "feedback". Our recent opening lines are listed above: do not repeat any of their phrases, even short thank-you words, so this reply does not sound copied.
2. Mention one or two specific things the reviewer wrote about, in your own words. Use only what is in the review and keep it accurate: same order, same timing, same people. Do not say how they felt unless they said it. Do not add anything about their visit, their job, what they bought or what happened that they did not write.
3. If the reviewer praised a team member who is on the list above, say you will pass their thanks on to that person by first name. Do not name anyone else.
4. Do not reveal anything private: no prices, addresses, dates, health details, account or job details, even if you could guess them.
5. Do not offer a discount, gift, reward or anything else, and do not ask them to leave more reviews, refer friends or share anything.
6. Write like a person, not an advert: do not list our services, do not add our town or keywords, and use the business name only in the sign-off.
7. Do not invent facts, numbers, promises, guarantees, policies or awards, and do not say we are the best at anything.
8. Write as "we" (the business) speaking to the reviewer as "you". Never call the business "they". End with one short, warm sentence that fits what they wrote (for example, if they said they will be back, say you look forward to seeing them), without promising anything.
9. Keep it to 2 to 4 short sentences and 25 to 60 words before the sign-off. No emojis, no hashtags, at most one exclamation mark.
10. Finish with a sign-off line in this exact format: {{SIGNER}}, {{BUSINESS_NAME}}
Return only the finished reply, including the sign-off line, and nothing else.What to fill in
- {{BUSINESS_NAME}}
- Your business name exactly as it appears on Google.
- {{BUSINESS_TYPE}}
- What you do and where, in a few words. The AI is told not to repeat your town or services in the reply.
- {{SIGNER}}
- First name and role of the person signing the reply.
- {{TONE}}
- Two or three words for how you sound, e.g. warm, down-to-earth, brief.
- {{STAFF_OK}}
- First names of team members who are happy to be named in public replies, or "none".
- {{DO_NOT_MENTION}}
- Anything that must stay out of a public reply, e.g. prices, job or order details, addresses.
- {{RECENT_OPENINGS}}
- The first lines of your last two or three review replies, so this one does not start the same way.
- {{REVIEWER_NAME}}
- The reviewer's name as Google shows it. The AI is told to use the first name only.
- {{STARS}}
- The star rating they left.
- {{REVIEW_TEXT}}
- The review, pasted word for word.
Real example output
Fictional sample business. Riverbend Auto Care, Gina, Marcus and the reviewer Priya S. are fictional. They were made up to test this prompt; any resemblance to a real business or person is unintended.
Show the inputs used for Riverbend Auto Care (fictional)
- BUSINESS_NAME
- Riverbend Auto Care
- BUSINESS_TYPE
- independent car repair shop in Nashua, NH
- SIGNER
- Gina, owner
- TONE
- warm, down-to-earth, brief
- STAFF_OK
- Marcus
- DO_NOT_MENTION
- prices, the bill or the quote, the car's details beyond what the reviewer wrote, what was repaired
- RECENT_OPENINGS
- "Thanks so much, Tom." / "We really appreciate you taking the time to write this, Linda."
- REVIEWER_NAME
- Priya S.
- STARS
- 5 stars
- REVIEW_TEXT
- My check engine light came on two days before a road trip and they fit me in right away. Marcus explained the problem in plain English, showed me the old part, and had the car ready the same afternoon. The final bill matched the quote exactly. I'll be back for oil changes.
What the model returned
Priya, thanks for your kind words about Marcus and how he explained the issue in plain English. You mentioned he showed you the old part and had your car ready the same afternoon—great to hear that worked well for you. We’ll pass that thanks along to Marcus. Looking forward to seeing you again for your oil changes. Gina, owner, Riverbend Auto Care
Model: Qwen3-4B-Instruct-2507 (Q4_K_M GGUF), llama.cpp server b11853, run locally, temperature 0.4. Tested on . Edits: None. The words are exactly as the model returned them, apart from trimming spaces at the ends of lines. All 8 runs with this prompt and these inputs passed the automatic checks (first name only, a thank-you up front, a detail from the review, thanks passed to Marcus, length, sign-off, no prices, offers, review requests, town or keywords). On a manual read 5 of the 8 were fine to post as written; 2 had a muddled sentence ("You were fit in right away and had your car ready", which reads as if the customer fixed the car) and 1 praised its own "great service". 3 of the 5 good ones also added "that's exactly what we aim for", which is harmless but generic. With the same review every run opened with nearly the same line, so the wording varies more between different reviews than between runs. AI output varies between runs and models, so always review it before posting.
Check before you post
- Everything the reply says happened is in the review itself, in the same order and with the same people. Cross out anything the AI added about their visit or how they felt.
- Only first names appear: the reviewer's (no last name or initial) and any team member who has agreed to be named.
- No prices, order or job details, addresses or anything else the reviewer did not post themselves.
- No discount, gift or reward, and no request to leave more reviews, refer friends or share the post.
- It reads like a person: the business name appears only in the sign-off, and the reply does not list your services or your town.
- Its opening is not the same as your last few replies. If it is, change the first few words before you post.
- Health, dental, therapy and other regulated practices: do not use this as it is. In our test on a fictional dental review, the small model said "your visit" or "welcome you back" in every run, even when told not to confirm the reviewer was a patient. Follow your own privacy rules for any public reply.
Get the full set of local business prompts.
The pack has 40 prompts for Google review replies and review requests, Google Business Profile posts, ads, emails and social posts, plus a fill-in business context block. One-time purchase, $24.
Related tools
Want to run prompts like this from your own code? Bifrost and Opper in the Digidle directory put many models behind one API. This example ran on a small open model on an ordinary computer; llmspeed previews local model speed before you buy hardware.