Yes, an AI teammate can prepare a review request after each completed cleaning visit, within rules the owner establishes in advance and reviews regularly. The owner decides which customers receive a request, what the message says, and when a person must take over the conversation personally. Hey Button runs on OpenClaw, an open automation framework, with models such as Claude and GPT answering requests through it.
Owners across Northeast Indiana, from Fort Wayne to Auburn, face the same timing problem. The work is good, but the request for feedback is frequently never sent, despite its obvious value. The sections below explain why that happens and how a clear routine corrects it.
Why do cleaning businesses lose reviews they could have earned?
Most cleaning businesses lose reviews because feedback requests arrive late or never arrive at all. The crew departs, the invoice is issued, and the owner intends to request feedback later in the week, when there is more time.
That later moment seldom arrives, because administrative work accumulates faster than any single person can clear it. A customer who was pleased on Tuesday has often forgotten the particulars of the visit by Friday afternoon.
A request arriving a full week later also reads as a formal obligation, so timing considerably shapes how the message is received.
A dependable routine corrects this pattern without requiring any staff member to remember anything at all. Each completed visit triggers exactly one request, and no staff member needs to keep track of who has been contacted. The owner establishes the governing rules once, and the work that follows becomes predictable and easy to audit at any later point.
Repetitive, rule-based work of this kind is precisely the category that an AI teammate handles with considerable reliability. Reliability and consistency matter far more than creativity in this setting.
Smaller businesses face an added constraint, since an owner who cleans by day cannot write a personal note to every client. A teammate can draft those notes in advance and place them in a queue. Each one then waits for approval until the wording reflects the owner's voice and standards of professional courtesy.
What should a review request actually say?
A good request names the service performed, thanks the customer by first name, and offers one direct link to the review page. It stays within three or four sentences, because longer requests tend to be skimmed or ignored entirely.
It does not ask for five stars, since that wording pressures the customer toward a particular answer. It does not offer a discount, a complimentary visit, or any other inducement in exchange for a review.
Google's review policies prohibit incentivized reviews, and the same restraint protects the business from a dispute that could surface months later.
The following example illustrates a suitable message for a recurring residential clean:
"Hi Dana, thanks for having us out today. If the kitchen and bathrooms met your standard, we would be grateful for a short review. Here is the link to the review page: [link]. If anything missed the mark, reply to this message and the owner will make it right."
The final sentence performs important work that is easy to overlook. It gives a dissatisfied customer a private channel, so a complaint reaches the business before it reaches a public page. The approach remains honest, because it does not prevent anyone from posting an accurate opinion about the service delivered.
When is the right time to ask?
Send the request within one business day of the visit, since prompt contact produces more reliable recall. For a morning appointment, the most suitable window falls in the same afternoon or the following morning. For an evening appointment, the message should wait until the next morning. Late-night messages tend to irritate recipients and measurably reduce the likelihood of receiving any response whatsoever.
If the owner defines quiet hours, the teammate should observe them without exception, regardless of how unusual the schedule appears on a particular day.
Recurring clients require a considerably different rhythm, one that protects their patience. A request after every weekly visit quickly becomes background noise, and people learn to ignore repetitive messages from a business they already trust. A better pattern asks after the first visit and again after a special engagement, such as a move-out or post-renovation clean.
For ongoing customers, the teammate should ask no more than once every three months. Document these timing rules during setup, so the teammate applies them identically rather than improvising from incomplete information.
How does the teammate decide whom to ask?
The teammate selects only the visits marked complete in your records, using the timestamps the scheduling system stores. It then excludes customers contacted within the previous ninety days, anyone who has opted out, and anyone with a formal complaint on file. The quality of those records therefore determines the quality of every request that follows.
Every completed job should appear in a booking sheet, a shared calendar, or a scheduling tool that the teammate is permitted to read. Errors in that source cause most of the early problems that owners encounter. The teammate may request feedback from people who never received service, or contact the same person twice within a short period, which annoys customers.
Before connecting any system, spend two full weeks carefully cleaning and verifying the customer list. Confirm that cancelled appointments carry a cancelled status, and that rescheduled visits carry a single, final status.
Crew members add valuable and frequently overlooked information about each property. Cleaners observe the actual condition of a home, notice damaged items, and hear comments that customers never write down. Give the crew lead a straightforward method for marking a visit "do not request." A single checkbox on the completion form is usually enough, and a flagged visit never reaches a customer who might respond with criticism.
Consider a simple, hypothetical calculation that represents no reported result, and suppose a crew completes forty visits in one month. Six of those customers received a request within the preceding ninety days, and one has opted out. The remaining thirty-three customers would each receive a single request, scheduled for the next business day. The arithmetic is straightforward, and it demonstrates why an accurate list matters more than any particular wording.
What should the teammate never do?
The teammate should never write, edit, or post a review on behalf of any person, because that practice is deceptive and violates platform policy. It should never offer an incentive, and it should never publicly reply to a review without explicit owner approval for that specific reply.
It should never limit requests to customers it expects to praise the business. Selective solicitation is the quickest route to losing a platform profile. Finally, it should never pressure anyone through repeated messages, since one request and one reminder represent a sensible maximum.
Public replies deserve particularly careful scrutiny, since each reply carries the business name for as long as the review remains visible. The teammate can draft those replies in a consistent professional tone, and an owner should still read and approve each one before publication. That habit prevents most of the embarrassing outcomes businesses describe after an automated system goes wrong.
How should a business respond to a negative review?
Call the dissatisfied customer on the same day, and reply publicly only after that conversation. The owner or crew lead should telephone or text the customer promptly, since a direct conversation resolves most disagreements quickly. That conversation should name the specific problem and apologize for it sincerely, without making excuses. It should also offer a return visit when the written policy supports one.
The public reply should then remain brief, factual, and courteous, and it should never argue with the customer on the record.
The teammate contributes by flagging the complaint quickly and drafting a first response that the owner can edit before anything is posted. It should not determine the remedy, promise a refund, or negotiate terms. Those decisions belong to a person who can stand behind them and answer follow-up questions with genuine authority and accountability.
How do you set up the routine and check its results?
Start with a written rule sheet that every team member can read and consult. List the trigger event, the waiting period, the approved message wording, the exclusion criteria, and the person responsible for complaints. Distribute that sheet to the teammate and the crew lead, so everyone works from one consistent set of rules rather than several competing memories.
Next, run a controlled test before any customer receives a live message from the system. Establish one practice job record, and ask one real customer who agrees to serve as a test recipient. Examine the message, the timing, and the resulting log entry carefully. Revise anything that reads awkwardly, because a first impression formed by a poorly phrased message is difficult to reverse.
During the first two weeks, retain the teammate in draft-and-approve mode, and review every request before it leaves the business. If messages read naturally and the log stays accurate, the owner may move the simplest requests to direct sending. Complaints and first-time customers should remain under approval for a considerably longer period.
The access level determines which systems the teammate can reach, so confirm that setting deliberately before connecting any customer list. Our guide to read-only, draft-and-approve, and full access explains how each level behaves in practice.
The log should show the date of each visit and the date of each request. It should also show the delivery status of each message, the status of any reply, and any complaint routed to a responsible person. Review the log once each month, and compare the number of completed visits against the number of requests sent during the same period.
A large gap means that completed jobs are being missed or overlooked entirely. It may also mean that the customer list contains errors, or that the teammate is excluding people for reasons the owner did not intend.
Track one additional measure as well: the number of requests that ultimately produced a published review. That figure reveals whether the message performs well with real customers in your market. If the rate stays low for two consecutive months, change one variable at a time, such as the wording, timing, or link placement. Changing a single variable at a time makes the cause of any improvement clearly identifiable.
When does this approach not fit?
Automation does not suit every cleaning business, and an honest assessment prevents wasted effort. A business that completes a handful of large jobs each year may do better with a personal message from the owner after each engagement. A company without a reliable completed-job record cannot automate dependably until that record exists and has been verified against the calendar.
A business that intends to solicit reviews only from its most satisfied clients should avoid automation entirely. That practice conflicts with platform rules and erodes the trust on which the business depends.
Privacy deserves careful and deliberate attention from the very beginning of any automation project. Customer names, addresses, and telephone numbers are sensitive information, so the owner should carefully limit what the teammate can access to the minimum required. The teammate does not need payment details, door codes, alarm information, or any other security data about the customer's home.
Sound review habits take several weeks to establish thoroughly, and patience and disciplined oversight both matter considerably. Begin with a modest scope, study the log carefully, and allow the process to earn confidence before expanding it.
Appointment reminders and follow-up messages are other tasks that a teammate can handle competently. The same access principles govern each of them, so a decision made for one routine usually informs the rest of the setup.
A cleaning company that asks every customer after every visit, with a clear route for complaints, typically accumulates honest reviews steadily over time. An AI teammate keeps that discipline running quietly in the background while the crew concentrates on the work customers actually notice and value.
The owner keeps the decisions that carry real consequences, and the teammate ensures that small, repeatable steps are never inadvertently forgotten. If you run a cleaning business and want a teammate to handle review requests under your rules, start with the Hey Button setup questionnaire.
Frequently Asked Questions
- Does the teammate need access to payment or invoice records?
- No. A review request needs only a first name, the service category, the visit date, and an approved contact method. Payment records add risk without improving the message, so keep them outside the teammate's reach.
- What happens if a customer asks not to be contacted again?
- Honor the request immediately and record it on an opt-out list that the teammate checks before every send. The record should be permanent, and no later visit should trigger a message to that customer.
- Can the teammate reach customers who prefer text messages over email?
- Yes, if each customer's preferred channel is recorded in the booking system and approved by the owner. Never switch channels without a recorded preference, because an unexpected channel can feel intrusive.
- How many visits should a business complete before it automates requests?
- No fixed number applies. Confirm first that completed-job records have been accurate for at least two weeks, because reliable records matter more than volume when the process begins.
- How can an owner tell whether the process is working?
- Compare completed visits with requests sent each month, and track how many requests produce a published review. Change one variable at a time, and allow several weeks before judging any change.
Hey Button
Want a teammate that handles this for you?
Answer a few questions about your business and Hey Button will show you what an AI teammate would take off your plate, with you approving anything that goes out.




