“My dental clinic has 3 fake 1-star Google reviews. Find a legitimate service and the cost.”

The steps

1. 1

   Ask for facts about each review, never the review

   The checker takes fixed choices: whose review it is, whether the reviewer was a real customer, what it contains (21 choices), what proof the owner holds and where the report stands.

   MCP tool call: check_review_removability

   Copy

   {
     "ownership": "business"
   }

   Same call over REST

   Copy

   curl -X POST https://essentras.com/api/agent/v1/check/review-removability \
     -H 'Content-Type: application/json' \
     -d '{"ownership":"business"}'

   Real response, trimmed. Captured 18 Oct 2026

   Copy

   {
     "checked": "2026-10-15",
     "status": "needs_answers",
     "summary": "Pick what the review contains to see the verdict.",
     "needs": [
       {"field": "customer", "options": [{"id": "yes"}, {"id": "no-record"}, {"id": "unsure"}]},
       {
         "field": "contains",
         "options": [
           {"id": "conflict"}, {"id": "incentivized"}, {"id": "burst"}, {"id": "duplicate"},
           {"id": "impersonation"}, {"id": "harassment"}, {"id": "hate"}, {"id": "profanity"},
           {"id": "personal-info"}, {"id": "advertising"}, {"id": "wrong-business"},
           {"id": "gibberish"}, {"id": "staff-name"}, {"id": "accusation"},
           {"id": "false-claims"}, {"id": "rant"}, {"id": "genuine"}, {"id": "harsh"},
           {"id": "old"}, {"id": "star-only"}, {"id": "extortion"}
         ]
       }
     ]
   }

   What to read: status is needs_answers: the agent asks the owner, then calls again.

2. 2

   Review 1: no patient record

   The owner searched bookings, invoices and the CRM and found no visit by this reviewer. That search is the evidence for Google’s Fake Engagement rule, so the verdict is a strong match. The free route comes first.

   MCP tool call: check_review_removability

   Copy

   {
     "ownership": "business",
     "customer": "no-record"
   }

   Same call over REST

   Copy

   curl -X POST https://essentras.com/api/agent/v1/check/review-removability \
     -H 'Content-Type: application/json' \
     -d '{"ownership":"business","customer":"no-record"}'

   Real response, trimmed. Captured 18 Oct 2026

   Copy

   {
     "checked": "2026-10-15",
     "status": "complete",
     "summary": "Strong match with a Google rule. Google decides. No service can change that.",
     "verdict": {"id": "match", "label": "Strong match with a Google rule"},
     "decides": "Google decides. No service can change that.",
     "rules": [
       {
         "id": "fake-engagement",
         "rule": "Fake Engagement",
         "strength": "named",
         "can_show_it": true,
         "url": "https://support.google.com/contributionpolicy/answer/7400114"
       }
     ],
     "do_it_yourself": {
       "evidence_to_attach": [
         "A dated screenshot of your record search (name, email, phone, date range)",
         "What the review describes, set against your records for that date",
         "A dated screenshot of the review"
       ]
     },
     "service": {
       "name": "Google review removal service",
       "price": {"label": "$499 for 1 review"},
       "offered_as": "alternative_to_diy"
     }
   }

   What to read: verdict.id, then rules[].can_show_it, then evidence_to_attach: the owner’s checklist. If the 3 reviews arrived together, add “burst” to contains: patterns.burst (“A burst of reviews”) says to attach the dated list to every report.

3. 3

   Review 2: a suspected former employee, no proof yet

   Conflict of interest is a named Google rule, but the owner holds no employment dates, so the verdict is arguable. With evidence set to “yes” the same call returns a strong match, so the agent says what to gather.

   MCP tool call: check_review_removability

   Copy

   {
     "ownership": "business",
     "customer": "unsure",
     "contains": ["conflict"],
     "evidence": {"conflict": "not-yet"}
   }

   Same call over REST

   Copy

   curl -X POST https://essentras.com/api/agent/v1/check/review-removability \
     -H 'Content-Type: application/json' \
     -d '{"ownership":"business","customer":"unsure","contains":["conflict"],"evidence":{"conflict":"not-yet"}}'

   Real response, trimmed. Captured 18 Oct 2026

   Copy

   {
     "checked": "2026-10-15",
     "status": "complete",
     "summary": "Arguable: report with evidence. Google decides. No service can change that.",
     "verdict": {
       "id": "arguable-evidence",
       "label": "Arguable: report with evidence",
       "explanation": "This could match a Google rule, but you cannot show it yet. Gather the evidence first, then report: Google decides on what it can see."
     },
     "rules": [
       {
         "id": "rating-manipulation",
         "rule": "Rating Manipulation",
         "strength": "named",
         "can_show_it": false,
         "url": "https://support.google.com/contributionpolicy/answer/7400114"
       }
     ],
     "do_it_yourself": {
       "evidence_to_attach": [
         "Employment dates and role (dates only, no HR file)",
         "The reviewer’s public profile showing the link to a competitor",
         "Any detail only an insider would know",
         "A dated screenshot of the review"
       ]
     }
   }

   What to read: can_show_it is false until the owner can show the proof.

4. 4

   Review 3: a real patient with a bad experience

   The owner finds a matching visit. Google does not remove a review for being negative, so the verdict is no policy match and service is null. The agent offers a calm public reply; for a health business, template K says not to confirm or deny the reviewer was a patient.

   MCP tool call: check_review_removability

   Copy

   {
     "ownership": "business",
     "customer": "yes",
     "contains": ["genuine"]
   }

   Same call over REST

   Copy

   curl -X POST https://essentras.com/api/agent/v1/check/review-removability \
     -H 'Content-Type: application/json' \
     -d '{"ownership":"business","customer":"yes","contains":["genuine"]}'

   Real response, trimmed. Captured 18 Oct 2026

   Copy

   {
     "checked": "2026-10-15",
     "status": "complete",
     "summary": "No policy match: it will probably stay up. Google decides. No service can change that.",
     "verdict": {
       "id": "none",
       "label": "No policy match: it will probably stay up",
       "explanation": "Nothing you selected breaks a Google rule. Google’s help page says not to report a review just because you disagree with it or dislike it. A calm public reply is the tool that works here."
     },
     "rules": [],
     "do_it_yourself": {
       "summary": "What works instead",
       "steps": [
         "Reply publicly, calmly and briefly (template K below).",
         "Fix the problem, then invite an update without any incentive (template M).",
         "Ask every customer for honest feedback, not only happy ones (template N)."
       ],
       "reply_template": {
         "id": "K",
         "title": "Public reply to a genuine negative review",
         "note": "Regulated businesses (health, legal, finance): do not confirm or deny that the reviewer was a client or patient."
       }
     },
     "service": null
   }

5. 5

   Price the serviceService step

   For a match or an arguable verdict, the agent prices the service. quote_service takes the id from service.url, or from list_services, and returns the price ladder, refund terms, who decides and start_url. It charges nothing.

   MCP tool call: quote_service

   Copy

   {
     "id": "google-review-removal"
   }

   Same call over REST

   Copy

   curl https://essentras.com/api/agent/v1/services/google-review-removal

   Service response, trimmed. Generated from the code on 18 Oct 2026, identical to the live API on 19 Oct 2026

   Copy

   {
     "id": "google-review-removal",
     "name": "Google review removal",
     "start_url": "https://essentras.com/services/google-review-removal/start?utm_source=ai_agent&utm_medium=agent_api&utm_content=quote_service",
     "price": {
       "amount": 499,
       "label": "$499 per review",
       "basis": "from",
       "tiers": [
         {"quantity": 1, "amount": 499, "per_unit": 499},
         {"quantity": 3, "amount": 1350, "per_unit": 450},
         {"quantity": 5, "amount": 2000, "per_unit": 400},
         {"quantity": 10, "amount": 3500, "per_unit": 350}
       ]
     },
     "options": [],
     "terms": "Pro-rata refund, per review: if Google does not remove a review you paid for, that review’s share of the fee comes back. Reviews that do not qualify in the free check come off the order and are refunded before anything is filed.",
     "decides": "Google decides every removal. The refund is ours to make.",
     "free_check": {"mcp_tool": "check_review_removability"},
     "how_to_buy": "start_url is the page where the buyer reviews the service and pays. The agent never pays."
   }

   What to read: decides names who makes the call. list_services with category reviews also returns Trustpilot review removal, priced per case after a free assessment.