First decide which it is. A fake review breaks Google's rules (never a customer, a competitor, an ex-employee, spam or an extortion attempt) and should be reported. An unfair one is a real customer's harsh or partly wrong account, and it needs a calm reply instead. AI helps with both: checking your records, matching the review to a policy, and drafting the report and reply.
Be clear about what AI can't do. It can't prove who wrote a review, and it can't get one removed. Google removes reviews only when they break its content policies, and its help pages say plainly not to report a review just because you disagree with it, adding that Google doesn't get involved in conflicts between businesses and customers. The work is building a clear, evidenced case for the reviews that do break the rules, and replying well to the ones that don't.
Fake or unfair? Sorting the review in two minutes
Most reviews that feel fake are real customers having a bad day. Getting this wrong in either direction costs you: report a real complaint and you waste your one appeal; apologise to a fake and you make the complaint look true. Use the signals below as a first sort, then check your records.
| Signal | Likely category | First action |
|---|---|---|
| No matching customer, date or job in your records | Possibly fake | Check records under other names; then report |
| Several one-star reviews within hours from accounts with no history | Review attack | Screenshot everything; report the batch |
| A message demanding money to remove reviews | Extortion scam | Don't reply or pay; use Google's extortion report form |
| Details that fit a former employee or rejected applicant | Conflict of interest | Report under that policy; reply neutrally |
| Describes a different business's service or location | Wrong business | Reply to point it out; report as off-topic if clear |
| A real customer, with facts partly wrong or tone harsh | Unfair but genuine | Don't report; reply with one neutral fact |
| Insults, slurs, threats or personal details about staff | Policy breach, even if genuine | Report; short reply if any |
Step 1: Check the review against your own records
AI is good at comparing messy details: a first name, a rough date, a service mentioned, a complaint about a specific person. Paste the review and a short export from your booking or job system for the relevant weeks, with customer contact details removed. A prompt that works:
Below is a review and a list of our jobs from [date range].
Customer names are first name + initial only.
REVIEW: "[paste review, including reviewer display name and date]"
JOBS:
[paste: date, first name + initial, service, staff, notes]
1. List any jobs that could match this review, with reasons.
2. List any details in the review that don't match anything in our
records (services we don't offer, staff we don't have, dates).
3. Say how confident you are that the reviewer was a customer:
likely, possible or unlikely. Don't guess identities beyond the
records.An illustrative result for a laboratory that received a one-star review from "Mark T." complaining that his "blood test results were lost":
Illustrative AI output: "1. No jobs match 'Mark T.' in March. Two customers with first name Mark had water tests (Mark D., 4 March; Mark P., 19 March); neither involved blood tests. 2. The review mentions blood tests and a 'nurse at reception'. The job list shows only water, soil and food testing, and no clinical staff. 3. Unlikely to be a customer. The review may be intended for a different laboratory."
That's a useful answer, and it's only a first pass. The manager still searches the inbox and phone log for "Mark", because not every enquiry becomes a job. Nothing turned up, so the review goes into the "possibly fake or wrong business" pile, with the AI's comparison saved as part of the evidence.
Step 2: Match the review to a specific Google policy
Reports work better when they name the rule broken rather than saying "this is fake". Google's contribution policy includes these categories, among others:
- Fake engagement: content not based on a real experience, or that doesn't accurately represent the place.
- Conflict of interest: including current or former employment, or a contractual relationship. Competitors reviewing each other fall here too.
- Off-topic: content not based on an experience at that specific location, such as general rants or reviews meant for another business.
- Harassment, hate speech and personal information: abuse aimed at named staff, slurs, or someone's private details.
Ask the assistant to do the matching, with the policy text in front of it:
Here is a Google review and the relevant extracts from Google's
Maps content policy. Which policy category, if any, does it most
likely break? Quote the part of the review and the part of the
policy that match. If it doesn't clearly break any policy, say so.
REVIEW: "[paste]"
OUR RECORD CHECK: "[paste the Step 1 summary]"
POLICY EXTRACTS: [paste the categories from Google's policy page]The "if it doesn't clearly break any policy, say so" line matters. It gives the AI permission to tell you that a harsh review is simply a harsh review. In an illustrative run on a genuine but angry review ("Rip-off merchants, never again"), the assistant replied that it doesn't break any listed category, because strong opinion about a real visit is allowed. That's the answer you need before you waste a report.
Step 3: Report it, and keep an evidence pack
There are two routes for a normal report. From your Business Profile, open Read reviews, choose the report option next to the review and pick the reason. Or use Google's Reviews Management Tool, choose your business, select "Report a new review for removal", and pick the review and the reason. The tool also shows the status of every report: decision pending, reviewed with no policy violation, or escalated (check your email). If a review isn't removed, you get one appeal, and the appeal form lets you select up to 10 eligible reviews. So save it for your strongest cases, and group them.
Build the evidence pack as you go, in one document, because you may need it for an appeal, an extortion report or a solicitor. A filled-in illustrative example:
REVIEW EVIDENCE: "Mark T.", 1 star, posted 21 March 2026, 22:14
Link: [review link]
Screenshot: saved 22 March, 08:02 (review + reviewer profile page)
Record check: no job, enquiry or call from any "Mark T." Jan-Mar.
Review describes blood tests and a nurse; we offer neither.
Policy: Off-topic (not based on an experience at this location)
Reported: 22 March via Reviews Management Tool, reason: off-topic
Status: 25 March "Decision pending"; 29 March "Removed"
Public reply posted: 22 March (see text below)Screenshots matter because reviews and profiles can be deleted or edited. Capture the review, the reviewer's profile page showing their other reviews, and the time.
A laboratory's review wave and extortion email, day by day
The single fake review is the common case. The more frightening one is a wave. Here is an illustrative timeline for a small environmental testing laboratory with 62 reviews and a 4.8 rating.
- Tuesday evening: six one-star reviews arrive in under two hours. Three have no text; three say variations of "terrible service, avoid". None of the six accounts has any other reviews. The rating drops to 4.3.
- Wednesday, 9am: an email arrives from an unknown address: the reviews will be removed for a payment, and "more will follow" if the lab doesn't pay by Friday.
- Wednesday, 9.30am: the owner doesn't reply to the email. He screenshots the six reviews, the six reviewer profiles and the email with its full header, and runs the Step 1 prompt against the last three months of jobs. None of the six names matches a customer.
- Wednesday, 10.15am: he uses Google's merchant extortion report form, which asks for the business details, profile link, links to the suspicious reviews, anything known about the person (names, email addresses, phone numbers, social profiles), a timeline, and screenshots of the demands. The AI assistant turns his notes into a clean timeline for the form in about five minutes.
- Wednesday, 10.40am: he posts one short, identical public reply under each review (text below) and tells the team not to respond to any messages from the sender.
- The following week: in this illustration, Google removes the six reviews and the rating recovers. Google says investigations are confidential, so don't expect a detailed explanation either way.
Two things from Google's own guidance shaped those decisions. Its extortion help page says not to engage with or pay the people involved, and not to try to resolve it yourself by offering money or services. And in April 2026 Google announced that when it detects this kind of scam, it may remove the fake reviews, pause new reviews on the profile, alert the owner, and show customers a banner explaining why reviews are paused. If you see that banner on your own profile during an attack, it's protection, not punishment. Google also reported blocking or removing more than 292 million policy-violating reviews in 2025, so the automated systems catch a lot before you ever see it; your report fills the gaps. Review extortion is one of several scams now aimed at small firms, and the same keep-calm, keep-evidence approach applies to the others in spotting deepfake voice and video scams.
Step 4: Reply in public while the report is pending
A report can take days or weeks, and prospective customers will read the review meanwhile. A short, calm public reply tells them what you know without accusing anyone. Four templates to adapt; ask your assistant to fit them to the review, but keep the facts to what your records show.
SUSPECTED FAKE (no record of the customer):
"We've checked our records carefully and can't find any visit or
enquiry that matches this review. If we've missed something, please
call [name] on [number] so we can look into it properly."
WRONG BUSINESS:
"We think this review may be meant for another business: we don't
offer [service mentioned]. If you did visit us, please get in touch
with [name] and we'll look into it."
CONFLICT OF INTEREST (e.g. ex-employee):
"We can't find a customer record matching this review. We've reported
it to Google under its content policies. Anyone with a concern about
our service is welcome to contact [name] on [number]."
REVIEW ATTACK (post the same under each):
"We've received several reviews tonight from accounts we can't match
to any customer, and we've reported them to Google. Our genuine
customers' reviews are below; please call us with any questions."Notice what none of them does: name a suspect, mention the extortion email publicly, share customer details, or sound angry. The ex-employee template is deliberately silent about employment, because saying "this is a former employee we dismissed" publicly creates problems of its own.
An unfair but real review at a car repair garage
Most reviews that owners feel are fake are actually this kind. An illustrative example:
★☆☆☆☆ "Took my car in for a simple noise and they kept it three days. Total cowboys."
The records show the customer, the job, and a three-day wait. They also show why: the noise was a failing wheel bearing, the part was ordered the same morning and arrived on day three, and the customer was told by phone on day one. "Total cowboys" is harsh, but it's opinion about a real visit, which Google's policies allow. Reporting it would fail.
The owner's angry first draft, and the AI-assisted edit:
- Before: "This review is completely unfair. We explained on day one that the part had to be ordered, and you agreed. Calling us cowboys when we did exactly what we said is out of order."
- After: "Three days is a long time to be without your car, and we're sorry it dragged. The noise turned out to be a wheel bearing, and the part took until day three to arrive. We should have kept you better updated while we waited. Dan is happy to talk it through on the main number."
The "after" contains the same fact (the part had to be ordered) without arguing, and it owns something real (updates while waiting). A prospective customer reading both would trust the second garage more. For more complaint replies built this way, see the worked examples in replying to negative reviews with AI.
A misdirected review at a funeral director with a similar name
Businesses with common or similar names get misdirected reviews more often than they realise. An illustrative funeral director receives a two-star review complaining about "the chapel of rest on the high street being closed at weekends". It has no chapel on the high street; a similarly named firm a few streets away does.
Here the public reply comes first, because it helps the reviewer and readers straight away: "We think this may be meant for another funeral director with a similar name, as we don't have a chapel on the high street. If you did use our service, please call us so we can help." Then report it as off-topic. Take extra care with tone in a sector like this: the reviewer is probably grieving and has simply picked the wrong listing. AI drafts for this situation tend to sound slightly triumphant ("This review is clearly not about us"); soften them before posting.
An ex-employee posing as a customer at a care agency
An illustrative care agency receives a one-star review from a new account: "Staff are overworked and underpaid, management don't care about the clients, avoid." There's no matching client or family. The wording, the timing (a week after a carer left on bad terms) and details about internal rotas point to a former employee.
Google's conflict-of-interest policy covers current or former employment, so this is reportable. What the manager should not do is say so publicly or confront the former employee, because that turns a policy report into a personal dispute and may create an employment problem. The steps: screenshot, record check, report under conflict of interest, post the neutral template above, and note it in the evidence pack. If the review contains false statements about named individuals, keep the evidence and read the last section below.
Moves that make a fake review worse
Each of these feels satisfying for a moment and costs you later.
- Accusing someone by name in a public reply. Even if you're right, it can escalate into a public row, and if you're wrong, it's a serious problem of your own.
- Publishing customer details to prove a point. "Our records show you visited on 3 March for a brake job" confirms private information in public.
- Paying or negotiating with extortionists. Google's guidance is explicit, and payment usually invites the next demand.
- Asking friends, family or staff for positive reviews to bury it. That's fake engagement. Google can restrict profiles for it, including pausing new reviews, unpublishing existing ones and showing customers a warning.
- Reporting every negative review. It clogs your own tracking, and you get only one appeal. Report only what you can link to a named policy.
- Letting an auto-reply tool answer. A cheerful automated "Thanks for your feedback!" under a review attack looks absurd. Exclude one- and two-star reviews from any automation; whether to let AI auto-post review replies covers where to draw the line.
When a solicitor is the right next step
Occasionally a review goes beyond unfair into false statements of fact that damage the business: an accusation of a crime, or a claim that a named person did something they didn't. That can be defamation in many legal systems, but the rules, defences and costs vary a great deal, and legal action over a review can draw far more attention to it than the review ever had. So treat this as a question for a qualified solicitor, not something to decide from a template.
What you can do yourself is keep the evidence pack complete: dated screenshots, the record check, your report history and any messages. A solicitor can then tell you quickly whether there's anything worth pursuing, and whether a platform removal request or a letter is more proportionate than a claim. AI is useful for organising that pack into a clear timeline. It shouldn't draft legal letters or threats for you to send.
Most fake reviews end quietly: a report, a calm reply, and a steady flow of genuine reviews that make one odd one-star look like what it is. Ask every customer for a review as routine (automated review requests make that painless), reply to all of them, and keep your records tidy enough that the next time something doesn't fit, you can show it in ten minutes.
Fake and unfair review questions
How long does Google take to decide on a reported review?
There's no fixed time. Google's Reviews Management Tool shows the status of each report, such as decision pending, reviewed with no policy violation, or escalated with an email to follow. Check the tool rather than reporting the same review repeatedly, and keep your one-time appeal for reviews where you have clear evidence of a policy breach.
Can I see who wrote an anonymous Google review?
No. Google shows only the reviewer's display name and profile, and it won't give businesses contact details. You can often work out whether a reviewer was a customer by checking the name, date and details against your own records, but don't publish your guess. If the review is part of a legal dispute, a solicitor can advise whether a court process could require disclosure.
Will lots of genuine new reviews push a fake one out of sight?
Genuine reviews from real customers do reduce a single bad review's effect on your average and push it down the newest-first list. Ask every customer as a matter of routine. What you must never do is ask friends, family or staff to post reviews, or run a sudden campaign to bury one review, because that is the fake engagement Google penalises.
Do other review sites handle fake reviews the same way?
The broad pattern is similar: most sites have a report or flag option, publish content guidelines, and remove reviews only when those guidelines are broken, not because the business disagrees. The categories, evidence required and appeal routes differ, so read the site's own business help pages before reporting and match your report to their wording.
Further reads
- How to Use AI to Respond to Google Reviews — The everyday routine for answering the rest of your reviews.
- AI Review Management Tools for Small Businesses (2026) — Tools that alert you to review spikes across several sites.
- What Is Review Gating and Can It Get Your Business Penalised? — Stay on the right side of Google's rules while you rebuild.
- How to Optimise Your Google Business Profile With AI — Tighten the rest of your profile once the dust settles.
- Patient Reviews and AI Replies: Staying Within Confidentiality — Extra care when the reviewer's details are health information.
- How to Answer Restaurant Complaints With AI Without Escalating — Calm replies to genuine complaints in a busy sector.
- How to Reply to Hotel Reviews With AI Without Sounding Canned — Why AI review replies read as templated, and the voice sheet, reply shapes and prompt that make each one sound written by the person who runs the place.
- How Small Amazon Sellers Use AI for Listings and Reviews — How a small Amazon seller can use AI for listings under the 2026 title rules, analyse reviews for fixes, and grow reviews without breaking policy.
- 7 AI Mistakes Restaurant Owners Make With Bookings and Reviews — Seven ways restaurant AI goes wrong with tables and reviews, each with a real-looking example, what it costs and how to fix it.
- Getting More Reviews After Every Job With AI Follow-Ups — When to ask for a review after each kind of trade job, the AI prompt that makes each request personal, and the Google rules that catch firms out.
- What AI Can and Can't Do for Your Local SEO — Which local SEO jobs AI does well, which it can't touch, and an afternoon's worked example for a dry cleaner, mapped against Google's own ranking factors.
- Can You Use AI to Write Testimonials? Where the Legal Line Is — Where the legal line sits for AI and testimonials: an allowed-grey-never table, what the rules prohibit, an approved-edit workflow and a consent email.
- AI Tools and AI Development: The Complete 2026 Guide — the AI hub, including every tutorial in the AI-for-business series.
Sources: Google Business Profile Help (Report inappropriate reviews; Report negative review extortion scams; Business Profile restrictions for policy violations); Google Maps user-contributed content policy (prohibited and restricted content); Google's announcement 'New ways we're protecting businesses on Maps' (16 April 2026).