Back to Phantom Notes
Automation

How to Build a Google Review Alert System with AI and n8n

March 21, 20267 min readBy T.W. Ghost
n8nGoogle ReviewsAI AutomationBusinessReputation Management

Why This Matters

A 1-star Google review sits on your profile for everyone to see. The longer it goes unanswered, the more damage it does. Studies show that 53% of customers expect a business to respond to negative reviews within a week. 1 in 3 expect a response within 3 days.

Most businesses find out about bad reviews days later when someone happens to check. By then the customer is gone and potential customers have already seen the unanswered complaint.

The fix is simple: automate the detection and draft the response the moment a review drops.


What This Automation Does

Here is the flow at a high level:

  • Monitor - Check for new Google reviews on a schedule (every 15 minutes)
  • Analyze - AI reads each new review and scores the sentiment (positive, neutral, negative)
  • Alert - If the review is negative (1-2 stars), send an immediate Slack/Teams notification to the team
  • Draft - AI generates a professional, empathetic response tailored to the specific complaint
  • Queue - The draft response lands in your inbox or Slack for a human to review and post

The entire process takes less than 60 seconds from review to draft response in your hands.


The Architecture

ComponentToolRole
Schedulern8n Schedule TriggerPolls every 15 minutes
Data sourceGoogle My Business APIFetches new reviews
AI analysisClaude or ChatGPTSentiment scoring + response drafting
NotificationSlack or Microsoft TeamsAlerts the team instantly
StorageGoogle Sheets or AirtableLogs all reviews + responses for tracking

You do not need to be a developer to build this. n8n is visual, drag-and-drop workflow automation. If you can connect boxes with lines, you can build this.


Step 1: Connect to Google Reviews

Google My Business API gives you access to your reviews programmatically. You need:

  • A Google Cloud project with the Business Profile API enabled
  • OAuth2 credentials (n8n handles the auth flow for you)
  • Your Google Business location ID

In n8n, use the HTTP Request node to call the reviews endpoint. The API returns each review with the reviewer name, star rating, comment text, and timestamp.

Set the Schedule Trigger to run every 15 minutes. On each run, compare against the last review timestamp you stored to only process new reviews.


Step 2: Score Sentiment with AI

Not every review needs the same response. A 5-star review gets a quick thank you. A 3-star review gets a thoughtful acknowledgment. A 1-star review gets a carefully crafted response that addresses the specific complaint.

Feed each new review into an AI node with this prompt structure:

Role: You are a customer service manager for a professional business.

Task: Analyze this Google review and provide:

  • Sentiment score (positive, neutral, negative)
  • Key complaint topics (if any)
  • A draft response that is empathetic, professional, and addresses the specific concerns

Constraints: Keep the response under 150 words. Never be defensive. Always acknowledge the customer's experience. Offer to make it right with a specific next step.

The AI returns structured output you can route through the rest of the workflow.


Step 3: Route by Severity

Use an n8n If node to split the flow:

  • 1-2 stars (negative): Immediate Slack alert + AI draft response + log to spreadsheet
  • 3 stars (neutral): Slack notification + AI draft response
  • 4-5 stars (positive): Auto-log to spreadsheet + optional thank-you draft

The critical path is the negative review. When a 1-star review hits, your team should know within minutes, not days.


Step 4: Alert Your Team

Get the Weekly IT + AI Roundup

What changed this week in NinjaOne, ServiceNow, CrowdStrike, and AI. One email, every Monday.

No spam, unsubscribe anytime. Privacy Policy

The Slack message should include everything the team needs to act:

  • The reviewer's name
  • Star rating (with color coding)
  • The full review text
  • The AI-drafted response (ready to copy and paste)
  • A direct link to respond on Google

One glance at Slack and your team can post the response. No logging into Google Business, no hunting for the review, no staring at a blank text box trying to figure out what to say.


Step 5: Track Everything

Log every review to a Google Sheet or Airtable base with columns for:

  • Date and time
  • Reviewer name
  • Star rating
  • Review text
  • Sentiment score
  • AI draft response
  • Actual response posted (filled in manually)
  • Response time (calculated)

After a month, you have data on your average response time, most common complaints, and sentiment trends. That data drives real business decisions.


What the AI Response Looks Like

Here is a real example of what the AI drafts for a 1-star review:

Review: "Waited 45 minutes for service. Staff seemed uninterested. Will not be coming back."

AI Draft:

"Thank you for taking the time to share your experience. We sincerely apologize for the long wait and the impression our team gave. That is not the standard we hold ourselves to. We would love the opportunity to make this right. Please reach out to us directly at [contact] so we can address this personally. Your feedback helps us improve, and we hope to earn back your trust."

The draft gives your team a professional starting point. They can personalize it, add specific details, and post it in under 2 minutes.


Why Not Just Use Google's Built-In Notifications?

Google does notify you about new reviews. But:

  • The notification does not include a draft response
  • It does not score sentiment or prioritize urgent reviews
  • It does not log reviews for tracking and analysis
  • It does not route alerts to your team's Slack where they actually work
  • It does not give you response time metrics

The automation does all of this without anyone checking a dashboard or opening an app.


How This Scales

One location? This workflow handles it easily. Ten locations? Add a loop node in n8n that iterates through each location ID. The architecture is the same, you just process more data per run.

For MSPs and IT consultants managing multiple client businesses, this becomes a service you can offer. Set up the workflow once, clone it per client, and charge a monthly fee for reputation monitoring. One n8n instance handles all of them.


What This Costs

ComponentCost
n8n (self-hosted)Free
Google My Business APIFree
Claude API (per review)~$0.003 per review
SlackFree tier works
Total for 100 reviews/monthLess than $1

Compare that to reputation management software that charges $50-300/month for essentially the same thing, minus the AI-drafted responses.


Getting Started

If you are already running n8n, you can build this in under an hour. The key components are:

  • Schedule Trigger (every 15 min)
  • HTTP Request node (Google My Business API)
  • Function node (filter new reviews only)
  • AI node (Claude or ChatGPT for sentiment + draft)
  • If node (route by star rating)
  • Slack node (alert for negative reviews)
  • Google Sheets node (log everything)

The full n8n workflow template with ready-to-use prompts and node configurations is available in our n8n Administration playbook, included with a Pro membership.


*Want to find the right AI and automation tools for your business? Take the free quiz and get matched with a personalized learning track in 2 minutes.*