RecipeAgents / Guides
Agent recipe guideAutomate Google review responses with AI
Replying to every Google review is good practice and easy to fall behind on. The hard part isn't the five-star 'thanks so much', it's the two-star review that needs a calm, specific answer on a day you're busy. This recipe drafts both for you, in your voice, and sends the draft to you for approval. Nothing is posted until you say so.
It uses Google Business Profile through Zapier or Make, one AI step, and email or Slack for the approval. Setup takes about 30 minutes.
The prompt
Draft a reply to this {{stars}}-star review in a warm, plain voice. 5–4 stars: thank them, mention one specific thing they said. 1–3 stars: thank, acknowledge the specific issue, no excuses, offer a direct contact. Under 60 words. Never offer refunds or discounts. Review: {{text}}Step by step
- Trigger: a new review on your Google Business Profile, connected through Zapier or Make.
- Send the star rating and the review text to an AI step with the prompt above.
- Send the draft to yourself by email or Slack, with the original review and a link to it.
- Read it, edit if needed, and post it yourself or through the integration once you approve.
Why draft-first, not auto-post
A reply to a review is public and sits under your business name for years. A wrong tone, a promise you can't keep or a detail the model misread costs more than the minute it takes to read the draft. Keep the approval step on for every review at first. Some owners later post positive-review replies with a quick skim and keep a careful read for anything three stars or below; that's your call once you've seen a few weeks of drafts.
What the prompt is doing
The prompt splits reviews into two groups. For four and five stars it asks for thanks plus one specific thing the customer said, which is what stops every reply reading like the same template. For one to three stars it asks for thanks, an acknowledgement of the specific problem, no excuses, and a direct contact so the conversation can move somewhere private. The 60-word cap keeps replies readable on a phone.
The line 'Never offer refunds or discounts' matters most. Without it, models tend to smooth over complaints with an offer, and a public offer under a review is one you then have to honour. If you do want to make it right, do it in the private follow-up, not in the public reply.
An example, with invented text
Here's an illustration, not a real review. Input, two stars: 'Coffee was good but waited 25 minutes for a sandwich.' Draft: 'Thanks for telling us. Twenty-five minutes for a sandwich is too long, and we're sorry. Glad the coffee was good! If you're up for it, email us at hello@… and we'll make the next visit better.' It names the actual problem, doesn't argue, and moves the follow-up to email.
The short version of this recipe lives in the library: draft replies to Google reviews. If you'd rather build it as an agent in a chatbot builder, our sister site has a review and comment reply drafter template.
Mistakes to avoid
Posting without reading. Keep the approval step, especially for low ratings.
Repeating personal details from the review, or adding any about the customer. Reply about the experience, not the person.
Letting the reply promise refunds, discounts or fixes. The prompt bans offers for a reason.
Arguing with a review you think is unfair. Acknowledge, offer the direct contact, and take it offline.
Build it in Make, Zapier or n8n
The recipe is tool-agnostic. Every step maps to a standard building block: a trigger from the app where the data starts, an AI step that sends text to a model with the prompt, a parser that reads the JSON or text that comes back, and actions in the destination apps. Zapier is the quickest to get running for linear flows. Make makes branching and loops easier to see. n8n can be self-hosted if you want to keep data on your own infrastructure. Pick the one your team already pays for.
Test it before you trust it
Run the recipe on ten to twenty real examples before you switch it on for everything. Put the inputs and outputs side by side in a sheet and mark each one right or wrong. Where it's wrong, add one specific line to the prompt that addresses the pattern, then run the same examples again. Keep a human approval step on any action that is visible to customers or touches money until the error rate is one you'd accept from a new hire.
Watch running costs as well. Each AI step is billed by the model provider and each task or operation by the automation tool. Filtering inputs before the AI step, for example skipping newsletters or tiny files, is the easiest way to keep both down. Check current pricing on each vendor's site, since it changes often.
Before you use it
Every recipe on RecipeAgents follows the same shape: a trigger, the tools, the steps, and the exact prompt. Start with the human-approval version. Once you have watched a few dozen runs and trust the output, decide which steps can run on their own. Model and tool names change often, so check current pricing and features in each tool before you build.
Frequently asked questions
- Can AI reply to Google reviews automatically?
- It can draft them automatically. This recipe keeps a person approving each reply before it's posted, because the reply is public and stays under your business name.
- Zapier or Make for this?
- Either works. Use whichever one already connects to your Google Business Profile and to the email or Slack account you'll approve from.
- How should it handle negative reviews?
- Thank the customer, name the specific issue, make no excuses and offer a direct contact. The prompt never offers refunds or discounts in public.