Why reviews carry so much weight

People trust other customers more than they trust adverts. Reviews also feed local search rankings, and AI assistants increasingly quote them when someone asks which business to choose. A steady flow of real reviews works for you in all three places at once.

Six steps from claimed to compounding

  • Claim your profiles: Google Business Profile first, then the one or two platforms your customers actually use.
  • Complete them: hours, photos, services, the exact same name and address everywhere.
  • Ask every customer, at the moment they are happiest: after delivery, after a solved problem, after a compliment.
  • Make it one tap: a direct link in a message, a QR code on the receipt or the counter.
  • Reply to every review within a few days, good or bad.
  • Feed what you learn back into the business, then tell customers what you changed.

How to ask without being awkward

Keep it short and specific: "If you have a minute, a review on Google helps other people find us. Here's the link." Ask in person when you can and follow up once by message. Ask everyone, not just the customers you expect to be kind. A page of nothing but perfect scores looks less believable than one with a few honest fours.

What the law says about fake reviews

This is general information, not legal advice. Fake reviews have moved from bad manners to a legal risk.

  • United States: a Federal Trade Commission rule in force since October 2024 bans fake and AI-generated fake reviews, buying positive or negative reviews, undisclosed insider reviews and suppressing negative ones, with fines per violation.
  • United Kingdom: fake reviews and concealed incentives were banned under consumer law from April 2025.
  • European Union: traders must say whether and how they check that reviews come from real customers.
  • Platforms: most forbid review gating, which means sending only happy customers to the review page, and paying for reviews.

Things never to do

  • Write reviews for your own business, or have staff or family do it without saying who they are.
  • Reuse one customer's words under several names.
  • Offer discounts only in exchange for positive reviews.
  • Post fake negative reviews about a competitor.
  • Threaten customers who leave bad reviews.

Replying to a bad review

Reply for the next reader, not for the reviewer. Thank them, acknowledge the specific problem, say what you have done or will do, and offer a way to continue in private. Keep it under a hundred words and never argue point by point.

If a review breaks the platform's rules, for example because it is about a different business, contains abuse or comes from someone who was never a customer, use the platform's report process. A review being negative is not a reason for removal.

Putting reviews to work

  • Quote them on your site with the customer's permission.
  • Collect the three most repeated compliments and put them in your headline.
  • Collect the three most repeated complaints and fix them in that order.
  • Don't mark up your own reviews for star ratings in search unless the platform's rules allow it. Search engines ignore or penalise self-serving review markup.

Where AI helps, and where it must not

AI is good at the tedious parts: drafting a reply you then edit, summarising two hundred reviews into themes, translating a review written in another language, or rewriting a request message to sound less stiff.

It must not write the reviews. An AI-written review posted as a customer's is a fake review, and it is exactly what the new rules target.

In UncutGPT, an uncensored model is useful for one specific job here: paste in your reviews and ask for a blunt summary of what customers dislike. Mainstream assistants tend to soften it.