How to request and get more Google reviews with n8n and AI (no coding)

Most happy customers never leave a Google review for one simple reason: they forget as soon as they walk out the door. Unhappy customers, on the other hand, rarely forget. When you don't ask for reviews consistently, your Google Business Profile ends up shaped by the extremes.

In this guide you'll build an n8n workflow that waits for the right moment after an appointment or purchase, checks that the customer didn't have an unresolved issue, uses AI to write a short personal note mentioning the actual service, and sends your direct Google review link. Once reviews start coming in, pair it with the workflow for replying to Google reviews with n8n and AI.

What you're going to build

A daily four-step workflow backed by a "Reviews" Google Sheet:

Trigger: every afternoon it checks customers served yesterday (from Google Calendar or your sales/appointments sheet).
Safety filter: skips anyone with an open issue, cancellation, or a low score on your customer satisfaction survey, and never asks the same customer twice within 6 months.
Personalized AI message: a Basic LLM Chain writes a short, natural email that references the specific service and includes your direct 30-second Google review link.
One gentle follow-up: if after 5 days the row is still unmarked, it sends a single two-line reminder and stops.

What you need

An n8n instance (cloud or self-hosted).
Your direct Google Business Profile review link (the short https://g.page/r/.../review URL that opens the star rating box right away).
Google Sheets and Gmail (or your email provider) connected in n8n.
An API key from OpenAI or another AI provider with a monthly spend limit set.
A "Reviews" sheet with columns email, name, service, service_date, status (pending, sent, reminder_sent, reviewed, do_not_send), sent_at and notes.

Step 1: grab your short Google review link

Don't ask customers to search for your business on Google Maps: every extra tap cuts completion rates in half.

Open your Google Business Profile manager.
Click Ask for reviews.
Copy the short link (https://g.page/r/.../review). Save it — you'll pass it as a fixed variable in n8n.

Step 2: filter who should (and shouldn't) get a request

Create a workflow with a Schedule Trigger every day at 18:00. Read yesterday's completed appointments or orders with Google Sheets (Get Rows) or Google Calendar.

Before calling the AI, add a Filter (or IF) node that excludes:

Customers with status = do_not_send or an open support issue.
Anyone who already received a review request in the last 180 days.
Anyone who hasn't consented to follow-up emails or has opted out.

Important note on Google's policy: Google forbids offering incentives, gifts or discounts in exchange for reviews, and forbids deceptive review gating. What is good business practice is fixing problems privately first when someone tells you a service went wrong, rather than firing off an automated marketing email asking for a favour while their issue is still unresolved.

Step 3: write a message that doesn't read like a template

Identical copy-pasted blasts get ignored. Connect a Basic LLM Chain with your AI model and a Structured Output Parser so it returns clean JSON (subject and body). Use a prompt like this:

You work at [BUSINESS NAME], a [TYPE OF BUSINESS]. Write a short email to a customer we helped yesterday. Data: name {{ $json.name }}, service or product {{ $json.service }}, optional note {{ $json.notes }}. Rules: warm, grateful, natural tone; max 75 words; naturally mention the service they received without inventing details; include a single call to action with the exact link [GOOGLE_REVIEW_LINK] noting it takes under a minute; never offer discounts or gifts in exchange; don't ask for "5 stars", ask for their honest feedback. Return JSON only: {"subject": max 6 words, "body": email text}.

Connect your Gmail node next. For the first week, set it to Create Draft so you can eyeball drafts in 30 seconds before sending; once you're happy with the tone, switch it to Send.

Then update the row in "Reviews" with status = sent and today's timestamp in sent_at.

Step 4: send one reminder after 5 days (and stop there)

Plenty of people open the first email while busy, think "I'll do it tonight" and forget. A single polite nudge 5 days later often doubles responses without annoying anyone.

In the same workflow (or a second schedule at 11:00):

Filter rows where status = sent and sent_at is 5 or more days ago.
Check that the customer hasn't already left a review or replied.
Send a short two-line fixed email ("Hi {{ name }}, just in case my previous note got buried — here's the direct link if you have 30 seconds to share how your {{ service }} went: [LINK]. Thank you!").
Update status to reminder_sent. Never send a third message.

Step 5: close the loop when the review lands

When a customer posts their Google review:

Mark their row as reviewed so they don't get the 5-day reminder.
Run the Google review replies workflow with n8n and AI to draft a warm public reply thanking them and mentioning the service. Prompt replies show prospective customers that you're active and attentive.

And if a customer replies to your email with a complaint instead of leaving a review, route it straight to a person using the human handoff pattern in n8n.

Test it before turning it on for real customers

Add three test rows to your "Reviews" sheet using your own email: one normal row, one with status = do_not_send, and one with a sent_at date 6 days ago.
Run the workflow manually in n8n.
Confirm the do_not_send row is skipped, the normal row creates a natural draft with a working Google link, and the 6-day-old row moves to reminder_sent.
Once everything checks out, point it at your live appointments or orders.

Cost and limits

Cost: generating a short email costs a fraction of a cent per customer on your AI provider's API.
Google rules: never buy reviews, never offer discounts or freebies for reviews, and don't demand 5 stars — ask for honest feedback on their experience.
Channel: email is the simplest, lowest-cost way to start. If you use WhatsApp, initiating a conversation outside the 24-hour service window requires an approved Meta template via WhatsApp Cloud API.
Privacy: only message customers you have an existing relationship with, and honour opt-outs immediately.

Start with ready-to-import templates

If you want the building blocks ready without wiring everything from scratch, download the free WhatsApp agent or check out the full AgentKiosk pack: 8 tested n8n agents in English and Spanish — including Google review replies and Google Calendar bookings — that plug right into this review-request workflow.