Bluemoonkey / Community / Review Reply Manager

Review Reply Manager Blueprint

A complete architecture for an AI-powered review reply tool — sentiment analysis, automatic draft generation, tone matching, and human approval gates before posting. 11 nodes, ready to export to Cursor or Claude Code.

SaaS AI Automation 11 nodes
Table of contents
  1. What this app does
  2. Architecture diagram
  3. Node breakdown
  4. How it works end-to-end
  5. Cursor export prompt
  6. FAQ

What this app does

Review Reply Manager is for local business owners who get 20+ reviews per month on Google Business, Yelp, or Trustpilot and can't keep up with responding to all of them. The app fetches new reviews automatically, uses AI to classify sentiment (positive, neutral, negative), drafts an appropriate reply in the business's tone, and routes every draft through a human approval gate before posting.

Responding to reviews — especially negative ones — improves local SEO, customer trust, and retention. But most business owners either don't reply or reply with generic "thank you" messages. This app makes personalized replies scalable.

This blueprint maps out 11 nodes across frontend pages, backend APIs, AI prompts, a human checkpoint, a database, and a Google Business integration. It's designed for a single business location managing one or two review profiles.

Architecture diagram

Dashboard (Page)Reply Editor (Page)Settings (Page)
↕
Reviews APIReplies API
↕
Sentiment Analysis (AI)Reply Drafter (AI)
↕
Approval Gate (Checkpoint)
↕
Reviews DB (PostgreSQL)Review Fetcher (Cron)

Node breakdown

Frontend pages

Backend APIs

AI prompts

Human checkpoint

Database & services

How it works end-to-end

  1. Business owner connects their Google Business Profile in Settings via OAuth
  2. Every 2 hours, the Review Fetcher pulls new reviews from Google Business API
  3. Each new review is sent to Sentiment Analysis — classified as positive/neutral/negative with topics extracted
  4. Review appears on the Dashboard with sentiment badge and draft status
  5. If auto-draft is enabled, Reply Drafter generates a draft reply automatically (for negative reviews, the business owner must manually trigger drafting)
  6. Business owner opens the Reply Editor, sees the review alongside the AI draft
  7. At the Approval Gate, they review, edit, or regenerate the draft
  8. Once approved, the reply is posted to Google Business via the Replies API
  9. Review status updates to "posted" on the Dashboard
  10. Reply rate and average response time metrics update on the Dashboard

Negative reviews are the highest-stakes replies a business writes. A bad response can go viral. The Approval Gate ensures a human reviews every reply — especially for negative reviews where tone and empathy matter most. The AI drafts, the human decides.

Cursor export prompt

Open this blueprint in Bluemoonkey and export it, or copy the prompt below:

Ready to paste 
Build this app exactly as described.
Start with database schema and auth, then APIs, then frontend pages.

## Dashboard (Frontend Page)
List of recent reviews with sentiment badge (green/amber/red), star rating,
review text, draft status (no draft / draft ready / approved / posted).
Filter by sentiment, platform, date. Reply rate metric.

## Reply Editor (Frontend Page)
Side-by-side view of original review and AI-drafted reply.
Edit draft, adjust tone. Approve and post, or reject and regenerate.
Shows sentiment analysis breakdown for context.

## Settings (Frontend Page)
Connect Google Business Profile via OAuth. Set business name and industry.
Configure reply tone (professional, warm, casual). Set fetch interval.
Enable/disable auto-draft for positive reviews (negative always manual).

## Reviews API (REST Endpoint)
- GET /api/reviews — paginated list with sentiment and reply status
- GET /api/reviews/:id — single review with full text, sentiment, draft
- POST /api/reviews/sync — manual trigger to fetch new reviews

## Replies API (REST Endpoint)
- POST /api/replies/draft — generate AI draft for a review
- PATCH /api/replies/:id — edit draft text, approve, or reject
- POST /api/replies/:id/post — publish approved reply to Google Business
- GET /api/replies/stats — reply rate, avg response time, sentiment distribution

## Reviews DB (PostgreSQL)
- reviews: id, business_id, platform, reviewer_name, review_text, star_rating, sentiment, sentiment_confidence, topics, fetched_at
- replies: id, review_id, draft_text, final_text, status, approved_by, approved_at, posted_at
- businesses: id, name, industry, google_business_id, tone_setting, created_at
- users: id, email, business_id, created_at

## Sentiment Analysis (AI Prompt)
Input: review text and star rating
Task: classify as positive/neutral/negative with confidence score,
extract key topics (service, product, staff, pricing, location)
Output: sentiment label, confidence (0-1), topics array,
one-line summary of the review's core message

## Reply Drafter (AI Prompt)
Input: review text, sentiment, topics, business name, industry,
configured tone, star rating
Task: draft personalized reply addressing specific points in the review,
matching business tone, thanking positive reviewers,
offering resolution for negative reviews
Output: draft reply text under 500 characters (Google Business limit)

## Approval Gate (Human Checkpoint)
Every AI-drafted reply must be reviewed by a human in the Reply Editor
before posting. Edit text, adjust tone, regenerate, or reject.
No reply is ever posted automatically — even for positive reviews.

## Review Fetcher (Cron Job)
Runs every 2 hours. Fetches new reviews from Google Business Profile API.
For each new review, triggers Sentiment Analysis and stores result.
If auto-draft enabled for that sentiment type, triggers Reply Drafter.

For more prompt templates, see our Cursor prompt template or Claude Code prompt template. New to planning? Read how to plan an app before coding.

Fork this blueprint

Open the free canvas, import this architecture, and customize it for your review management workflow.

Open Canvas — Free →
AI App Blueprint Template ← AI Support Triage Newsletter Autopilot → All community blueprints →

FAQ

What is a review reply manager?

A review reply manager monitors customer reviews on platforms like Google Business, uses AI to classify sentiment (positive, neutral, negative), drafts appropriate replies, and routes them through a human approval gate before posting. It helps local businesses respond to reviews quickly and professionally.

Can I export this blueprint to Cursor?

Yes. Open this blueprint in Bluemoonkey's free canvas and export it as a structured AI prompt formatted for Cursor's Agent mode. Paste it into Cursor, add "Build this app exactly as described," and it generates the full project.

Does it auto-post replies without human review?

No. Every AI-drafted reply goes through an Approval Gate where a human reviews, edits, or rejects it before posting. This is intentional — AI-generated replies to customer reviews can be off-brand or insensitive, especially for negative reviews where tone and empathy are critical.

Does it support platforms other than Google Business?

The blueprint is designed for Google Business Profile, but the Reviews API and Review Fetcher can be extended to support Yelp, Trustpilot, or Facebook Reviews by adding platform-specific adapters. The sentiment analysis and reply drafting work on any review text.