Home · Business · Customer Service · Complaints & escalations

YES

As of 13 August 2026, AI can respond to a negative Google review.

Most people should hand this to a purpose-built tool.

Can you do it?

5 minutesto a draft.

15 minutesto something you’d act on.

Cost, all in£0

Skill neededchat-fluent

Who has to check ityou

What the alternative costsBotpress is an open platform for building LLM chatbots and agents, so it is an alternative when you need a managed customer-service workflow rather than one drafted reply.

If this goes wrong: the reply exposes private information, escalates the dispute or makes a public admission that damages trust in your business.

What to actually do

  1. Use a tool built for this

    The route this page recommends

  2. Do it yourself

    Second choice

    A chat interface, chat-fluent skill, and roughly 15 minutes until you can act on the result.

    How to actually do it

    1. Open the review in your Google Business Profile and copy its text into a working document, removing the reviewer’s personal details and any private order information.
    2. Gather the customer’s confirmed transaction or visit facts, the relevant refund or complaints policy, and the contact route you want customers to use for private follow-up.
    3. Paste the review, verified facts, policy and contact route into the prompt, then ask the chatbot to produce the public reply.
    4. Compare every factual statement in the draft with your records and remove anything you cannot confirm, including details that would identify the customer.
    5. Check that the reply acknowledges the experience, avoids arguing or admitting liability, contains no confidential information and gives a real private contact route.
    6. If the review raises serious allegations or a possible legal, safety, discrimination, fraud or data issue, send the draft and records to the responsible manager for a decision; otherwise paste the approved reply into Google Business Profile and publish it.

    Prompt

    Draft one public reply to this negative Google review for my UK business.

Review:
[PASTE THE REVIEW]

Verified facts about the customer’s visit or order:
[PASTE ONLY FACTS YOU CAN CONFIRM]

Our relevant policy or service terms:
[PASTE THE RELEVANT POLICY, IF ANY]

Preferred contact route for resolving the issue privately:
[EMAIL ADDRESS, PHONE NUMBER OR CONTACT PAGE]

Write in a calm, human and professional British tone. Acknowledge the customer’s experience without automatically admitting liability. Do not argue, blame the customer, repeat personal data, reveal order details publicly, invent facts, promise a refund or compensation, or make claims about the customer’s motives. Do not mention internal processes. Keep it concise, apologise for the experience where appropriate, state only the verified facts that are safe to publish, and invite the customer to contact us privately so we can investigate. If the review alleges serious harm, discrimination, fraud, a data breach, a regulatory issue or a legal dispute, do not draft a final publishable reply: flag the issue for a human manager and list the facts that need checking instead. Provide the draft first, followed by a short list of any factual statements I must verify before publishing.

    Open it prefilled in ChatGPT or Claude, or copy it into Gemini, which takes no prefill link.

  3. Hand it to a person

    The distant third

    A person who owns the outcome does this end to end, worth it when the failure is dear.

What it gets wrong

  • AI cannot know whether the reviewer’s account is accurate unless you provide and verify the underlying records.
  • AI cannot decide whether a public response amounts to an admission of liability or creates a regulatory or legal risk.
  • AI cannot repair the customer relationship or conduct the private investigation that may be needed after the reply.
  • AI cannot take responsibility for the business’s decision to publish the response.

Even on a YES, the friction has a name: stakes of error, judgement under ambiguity and relationship.

How we scored this

Five axes, each scored nought to two by hand: ten means AI carries the task cleanly, and the thresholds that turn a total into YES, PARTLY or NO are published in the methodology. Each axis name links to its definition.

AxisScore (0–2)
Output2
Inputs2
Verification1
Liability1
Effort delta2
Total8 / 10

The methodology and its thresholds are published in full.

FAQ

Can AI respond to a negative Google review?
Yes. It can draft a calm response from the review and your verified business facts, but you must check it before publishing because your business remains responsible for the wording.
Can AI post a reply to a Google review for me?
A chatbot can write the reply, but you should publish it through your Google Business Profile after approval. Do not give an automated system permission to post replies without a human check.
What should I say in response to a bad Google review?
Acknowledge the customer’s experience, apologise where appropriate, state only facts you can confirm and invite them to contact you privately. Do not argue, reveal personal information, speculate about motives or promise an outcome you have not approved.
Is it safe to use AI to reply to a Google review?
It is generally suitable for a checked first draft, not for unsupervised publication. The main risks are exposing private information, repeating unverified claims, escalating the dispute or making an admission that harms your business.

Nearby answers

All Complaints & escalations rulings

Assessed by gpt-5.6-luna (gpt-5.6-luna) on 2026-08-13, second-checked by an independent model. Wrong somewhere? Email [email protected] and it gets re-checked.

The newsletter

AI news, new answers and product picks, straight to your inbox.