For operations and customer experience
Every review from every site becomes a signal your tools can act on. A low rating starts a workflow, a count crosses your line, and the right manager gets the task, through webhooks, Zapier, n8n, the MCP server or the API.
WhenGrades.com new low rating
ThenYour automation opens a taskA new 1- or 2-star review starts the workflow in n8n, Zapier or Make, a Grades.com review search counts that location's low ratings this month, and the third one opens a task and posts to Slack.
WithGrades.com complaint themes"Ready two hours late."
Every Monday your agent reads the week's complaint themes and tagged reviews, then writes the operations lead a short digest with each point tied to the reviews behind it.
WithGrades.com every locationOn the first of the month your agent compares every location's rating, reply rate and unanswered low ratings against the month before and sends the regional lead what changed.
Shown with n8n. Works the same with Zapier, Make or your own AI agent. Integrations →
Why review data
Google, Facebook, Yelp and the other sites you connect land in one place for each location, as they are found, so nobody checks five dashboards.
Low ratings, unanswered reviews and reply rate for any period, next to the period before, for each location and for the whole group.
A new review, a reply, a tag or a draft waiting for approval is an event your webhooks, Slack and automation tools can act on.
Every comparison here is per location. Grades.com does not rank or name individual people.
Each piece runs per location, in the dashboard, and through the MCP server and the API for your agent.
Webhooks and automation triggers
Every new review, reply, tag and agent draft is a signed webhook event. Zapier, Make and n8n subscribe with one key, and so can your own endpoint.
Ask your agent Which 1- and 2-star reviews came in at our Northside location this month?
locations_reviews
"type": "review.created",
"company": { "name": "Bayside Auto Care, Northside" },
"data": { "review": {
"site": "google",
"rating": 2,
"text": "Ready two hours late, nobody called."
} }
Slack alerts
Chat notifications route by rule: 1- and 2-star reviews from one location to that location's channel, reviews tagged as a refund request to billing, five-star reviews to the whole team.
Ask your agent Which of our Slack channels get low ratings, and for which locations?
chat_hooks_list
#northside-ops
Grades.comApp9:14 AM
New 2-star review on Google
Bayside Auto Care, Northside ★★★★★
"Car was ready two hours after the time they gave me, and nobody called."
Open in Grades.comMulti-location
For each location: new reviews, average rating, low ratings, unanswered reviews and reply rate for any period, against the one before. Sort by any of them, or let your agent do the comparing.
Ask your agent Which location had the most unanswered low ratings last month?
locations_compare
This month, by location
Low ratings: reviews of 1 to 3 stars in the period
Insights and review tags
Every review is tagged by topic and sentiment, and Insights counts the complaint themes for any period. On Premium, AI Reputation alerts also tell you when assistants start repeating a new complaint about you.
Ask your agent What did customers complain about most this month, compared with last?
review_insights
Complaints, last 30 days
Review Disputes and Review Agent
A 1- or 2-star review opens a case in Review Disputes, one per customer, so someone owns the follow-up. Review Agent drafts the public reply, and a person approves it before anything posts.
Ask your agent Which open cases have had no contact for two days?
recovery_cases
Car was ready two hours after the time they gave me, and nobody called.
Draft reply
Chris, two hours past the time we gave you, with no call, is not how a pickup should go. We now call every customer as soon as a job runs late. Our service manager would like to hear what happened: northside@baysideauto.example.
The data is already there
No exports to stitch together. Reviews from every site you connect, counted per location and tagged, are one call or one event away for your automations and your agent.
https://mcp.grades.com/mcphttps://app.grades.com/api/v1POST /v1/webhooksRead-only keys are enough for counting and reporting; add write access only where a workflow replies. No AI client? Ask Grades, the assistant in your dashboard, answers the same questions.
Zapier
Makelocations_summarylast monthNorthside
4.4 avgLakeview
4.8 avgHillcrest
4.7 avgWhat the operations team gets
Each one runs per location, in the dashboard and through the MCP server and the API.
| Deliverable | Product | MCP tools | Plans |
|---|---|---|---|
| Low ratings, unanswered reviews and reply rate per location, for any period | Multi-location | locations_summary, locations_compare, locations_reviews | Every plan |
| Complaint themes, sentiment and reviews tagged by topic | Insights | review_insights, review_tags_list | Every plan: Google on Starter, every site from Pro |
| A case for each unhappy customer, with its status and owner | Review Disputes | recovery_cases | Every plan |
| A reply draft for every review, approved by a person | Review Agent | agent_actions, agent_approve | 10 drafts a month on Starter, 50 per location on Pro, no cap on Premium |
| Alerts when AI assistants start repeating a new complaint | AI Reputation | ai_perception, alerts_list | Answers on every plan; alerts on Premium |
| Slack alerts by rule, signed webhooks and the REST API | Developers | chat_hooks_list | Slack alerts, MCP and API on every plan; webhooks on Pro and Premium |
Not as a built-in count yet, and we would rather say so. Slack alerts and webhooks fire on each low rating as it arrives, filtered by location. The count lives in your automation: a new-review trigger, a Grades.com search for that location's 1- and 2-star reviews this month, and your threshold. An agent can do the same from the per-location figures.
Zapier, Make and n8n have Grades.com triggers and actions, any service that accepts a signed webhook can receive events, and agents connect through the MCP server. Destinations are whatever those tools reach: your help desk, your task board, email or Slack.
No. Review pages and requests ask every customer the same way, whatever their experience, with no rating-first step and no incentives. A workflow acts on reviews already posted; it never decides who gets asked.
No. Every figure is per location, never per person. Tags describe topics such as timeliness or pricing, not names, and we do not build rankings of people.
No. Yelp reviews come through as a rating and a link: Slack messages, webhooks and the MCP server never carry their text, so a workflow built on them never hands it to your model.
14-day free trial of Pro, no card needed.