// the tuesday guide · one task a week, automated

Google Reviews: Ask at the Right Moment and Reply to Every One

2026 · 10 · 13 — ONE TASK A WEEK, AUTOMATEDReviews requested at the right moment, replies that start with you

When people look for a mechanic, a hairdresser, a restaurant or a tradesperson nearby, they check the stars on Google before they even open the website. Yet reviews are one of those jobs that always get pushed back: you remember to ask for them when you have time, which is almost never, and you reply whenever you get around to it. Meanwhile your happy customers leave without writing a word, and the angry one writes right away.

This guide takes the job apart piece by piece: how it works today, what you can automate and what you can't, which tools to use, a complete example and how to start in one afternoon. With one rule that governs everything else: you ask everyone, not just the happy ones, and never in exchange for anything.

How it works today: reviews that show up by chance

In most small businesses, it goes like this. The job is done, the customer says thanks, and someone thinks “I should ask them for a review.” If you're lucky, they say it out loud, and the customer answers “sure, as soon as I have a minute.” Then, back home, they'd have to search for the business on Google, find the right listing, find the button: almost nobody makes it to the end.

On the other side, the reviews that do come in get read whenever someone happens to open the profile. Replies get written one evening every so often, all at once, often with the same copy-pasted line. And the negative review gets discovered days later, when whoever reads it has already decided to go elsewhere.

The typical mistakes are three: asking too late, making it hard to leave a review, and replying late precisely to the criticism, which is what new customers read most carefully.

What gets automated and what stays with you

What gets automated is everything repetitive: the request that goes out when the job is closed, with the direct link to the review window; the alert when a new one comes in; the draft reply prepared by AI; the instant alert when a one- or two-star review arrives.

What stays with you is everything that takes judgment. You read and approve every public reply, because it carries your name. You handle negative reviews yourself, from start to finish: a machine can't tell whether the customer is right, whether there was a misunderstanding, or whether you need to make a phone call before replying in public. And you decide what to change in your work when the same complaints keep coming back.

Then there's a line that doesn't depend on you but on Google's rules, and the automation has to respect it by design: the request goes to every customer, not just the ones who seem satisfied, and nothing is offered in return.

With the tools you already have, and when you need something custom

The starting point is your Google Business Profile, the listing that shows up in Search and Maps. From there you create the link to ask for reviews, and you can turn that same link into a QR code for your counter, your receipt or your invoice. It's free, and anyone who has claimed their listing already has it.

Next, you need to know when a job is done. Almost always, you already know it somewhere: the work order closed in your management software, the order shipped in your online store, the appointment past in your calendar, the “delivered” row in your spreadsheet. That event is the trigger: when it happens, the message goes out.

The message can travel by email, with a sending tool connected to your management software or spreadsheet, or on WhatsApp. Careful here: WhatsApp Business, the free app, lets you save a quick reply with the text and link ready to go, but you send it yourself with a tap. Truly automatic sending on WhatsApp goes through Meta's official APIs, and that's already a small project. We explain it on the page about bots for Telegram and WhatsApp.

Finally, the replies. Your profile alerts you by email or notification when a new review comes in, and from there you can paste the text into an AI assistant to get a draft. To make reviews, drafts and alerts flow on their own, without copy-pasting, you need to connect to the Business Profile APIs instead, and Google grants access on request: that's usually the point where you need something custom, like the AI automation for businesses we build starting from the tools you already use.

A step-by-step example: from the returned car to the published reply

Take an auto repair shop with management software for its work orders. Here's the full loop.

1. The car is returned. At pickup, at the register, the work order is set to “closed.” Nobody has to remember anything else.

2. The message goes out that same evening. A short, signed message: “Thank you for choosing us for your car's service. If you have a minute, your opinion on Google helps us a lot: [link].” The link opens the review window directly. One message only, to every customer, with no pushy reminders: Google's rules also ask you not to pressure people.

3. The review comes in. The system reads it, logs it in a spreadsheet with date, stars and text, and asks AI for a draft reply: short, mentioning the specific job, and signed with the owner's name.

4. The owner approves. They get the draft on their phone, edit it if needed and publish it. Two minutes, not an evening.

5. If it's one or two stars, everything changes. No draft to approve in a hurry: an alert goes out right away, the owner checks the work order, calls the customer if needed, and only then replies in public. Calmly, with none of the customer's private details, owning the mistake if there was one and inviting them to get in touch directly: that's the same advice Google gives.

How to start in one afternoon, and the mistake to avoid

You don't need to start with the full automation. In one afternoon you can do four things:

First, create the review link and the QR code from your Business Profile, and put them where customers will see them: counter, receipt, email signature, invoice. Second, write the request message once, short and with your name on it, and save it as a quick reply in WhatsApp Business or as an email template. Third, turn on notifications for new reviews on your phone. Fourth, reply to the unanswered reviews from the last few months, starting with the negative ones.

From there you move on to the rest one piece at a time: connecting the message to the job being closed, then the AI drafts, then the alert for low ratings. If your starting point is management software that today can't tell you when a job is done, it's worth fixing that first, maybe with a custom web-based management system.

The mistake to avoid is the filter: first asking “from 1 to 5, how did it go?” and sending only the people who answer 4 or 5 to Google. It looks clever, but it's exactly the “selective soliciting” Google prohibits. The second mistake, almost as common, is letting AI publish replies on its own: twenty replies that sound like they came from the same machine get noticed, and a wrong reply to a complaint does more damage than the complaint itself.

In short

Ask for Google reviews right after the job, with the direct link to the review window, from every customer and with nothing in return: Google's rules prohibit both incentives and asking only satisfied customers. The request, the alerts and the draft replies can be automated; approving the replies and handling negative reviews stay with a person. You start in one afternoon with a link, a QR code, a well-written message and notifications turned on, and you connect the rest one piece at a time.

Sources: Google (Maps user-generated content policy: fake engagement) · Google (Business Profile: tips to get and reply to reviews)

Frequently asked questions