Review & Reputation Engine
An AI-powered Make.com workflow that monitors incoming reviews, auto-replies to positive ones, and escalates negatives to a human via Trello — zero manual inbox time.
The problem
Reviews were piling up with no one responding
For a service business, review response time is a trust signal. Positive reviews left unacknowledged feel like indifference. Negative ones left unaddressed can compound into a reputation problem.
The issue wasn't lack of intent — it was that responding to every review manually meant someone had to check Google and Facebook daily, write a thoughtful reply, and escalate complaints to the right person. That's easy to skip when the day gets busy.
The goal: respond to positives automatically within minutes, and make sure negatives never slip through without surfacing to someone who can act on them.
Make.com workflow
The full Make.com workflow — Google Sheets trigger → AI sentiment → Router → Gmail or Trello
How it works
New review logged to Google Sheets
Reviews from Google and Facebook are entered into a central Review Log sheet. Each row captures: timestamp, reviewer name, platform, review text, and star rating. The Make.com workflow polls the sheet every 15 minutes for new unprocessed rows.
AI classifies sentiment and drafts a reply
An HTTP node sends the review text to the AI API. It returns a sentiment label (Positive or Negative) and, for positive reviews, generates a warm, brand-appropriate reply. The classification and AI reply are written back into the sheet automatically.
Router splits positive vs. negative
An Make.com Router node checks the Sentiment field. Positive reviews go down one branch; anything negative (including neutral 3-star reviews) goes down the escalation branch. No ambiguous cases fall through the cracks.
Positive → Gmail confirmation sent, row updated
The AI-generated reply is sent via Gmail as a notification confirming the auto-reply was posted. The row in Google Sheets is updated to Replied.
Negative → Trello card created for human review
Negative reviews are never auto-replied to. Instead, a Trello card is created in the New Negative Reviews list with the reviewer name, platform, star rating, and the full review text. A human handles the response. The sheet row is updated to Escalated.
System in action
Review log before processing — new rows waiting to be picked up
After processing — AI Reply and Status columns filled in automatically
Trello Reputation Engine board — negative reviews surfaced for human action
Gmail notification confirming a positive review was auto-replied
Design decisions
Never auto-reply to negatives
A bad AI reply to a complaint is worse than silence. Any review with negative or neutral sentiment is routed to a human, every time.
Sheets as the single source of truth
Every review — positive or negative — is logged with its sentiment, AI reply, and final status. Full audit trail with no extra tooling needed.
15-minute polling window
Fast enough that positive reviews feel promptly acknowledged. Slow enough to avoid rate limits and keep the workflow stable under real daily volume.
Tools used
Want something like this built for your business?
Tell me where reviews are piling up and what you want off your plate. I'll map the pattern and tell you what's realistic.