Your team closes a deal in Medford, the file gets archived, and nobody asks for the review. Or someone asks everybody at once and the wrong person gets a request. Real estate review request automation fixes the first problem and can create the second. This post covers the step that prevents it: a queue where a person approves every Google review request before it sends. It is one piece of a wider real estate past client follow-up automation system for South Jersey teams. Below: what the operator sees on screen, who can approve, and what blocks a send. Where we do not know something, we say so.
Why the approval step exists
Most review automation is a trigger and a text. A CRM status flips to "Closed," and a message goes out. That works until the data is wrong. A closing that fell through can still carry a closed status. A phone number can belong to a seller's ex-spouse. An agent can have a heated email thread nobody else has seen.
A review request is public-facing. It goes out under your name, and the answer lands on a profile that buyers read before they call. A bad send is not a typo you fix in an hour. It can be an awkward reply you read for years.
There is also a policy reason. Google Business Profile Help says merchants may not discourage negative reviews or selectively solicit positive ones. It also bars incentives such as discounts or free goods in exchange for a review. That page was current when we read it in October 2026. Read it again before you build, because Google edits it.
The stakes are real. Google Business Profile Help says a pattern of violations can restrict a merchant's access, and then all profiles tied to that account are suspended.
So the queue has two jobs. It catches mistakes, and it stops the system from quietly deciding who gets asked based on how happy they seem. We come back to that second point below.
What the operator sees in the queue
Think of the queue as a weekly worklist, not an inbox. Each row is one past client with one proposed message. The operator can approve, hold, or skip. Skipping needs a reason from a short list.
Here is the layout we would scope for a small brokerage. It is a design, not a report on a shipped result.
| Field | Why it is on the screen |
|---|---|
| Client name and property address | The approver confirms the right household and the right deal. |
| Agent of record | Decides who is allowed to approve the row. |
| Side (buyer or seller) and closing date | Shows how long ago the deal closed and which story applies. |
| Last logged contact | A review request that follows a cold gap reads differently from one that follows a call. |
| Which Google profile link will be used | Agent, team and brokerage profiles differ. The wrong one sends the review to the wrong place. |
| Check results (pass or blocked) | Shows exactly which rule stopped a row. |
| The full message text | The approver reads what the client will read. |
Most fields on that list come from your CRM. If they are empty or inconsistent, the queue will show it immediately. That is why we start with the data. Our post on auditing your past-client list before the holidays covers the fields every follow-up system needs, and the review queue depends on the same ones.
Two design choices matter more than they look. First, the message text is visible in full, not behind a "preview" link. People approve what they can see. Second, blocked rows stay in the queue with the reason attached, instead of disappearing. A row that vanishes is a row nobody can question later.
Who can approve, and who cannot
Put the approval rule in writing before you build anything. Ours would have three roles.
- Agent of record. Approves requests for their own clients. They know whether the closing was smooth.
- Operations lead. Approves for any agent and clears holds. This is usually the transaction coordinator or office manager.
- Admin. Edits templates, check rules and the per-week cap. This should be one or two people, not the whole office.
An agent cannot approve a request for a client who belongs to another agent. Nobody can approve a row that has a hard block. And the person who edits a template should not be the only person who approves the sends that use it.
Every approval writes a log entry: who approved, when, the exact text sent, and which profile link was used. If a client complains, you can answer in a minute instead of guessing. If your brokerage has its own compliance contact, ask them whether they want to see that log. That is a question for them, not for us.
What if the agent never opens the queue? Rows older than a set number of days go to the operations lead, who can approve, hold or skip them. Pick the number with your team. We do not have a benchmark for it.
What blocks a send
A block is a rule the approver cannot click past. A warning is something the approver can override with a reason. Keep the hard blocks few and the warnings honest.
Data blocks: the record is not good enough to send
- The CRM status is not a confirmed closing. Pending, withdrawn and terminated deals never qualify.
- The closing date is blank or sits in the future.
- The email address is missing, malformed, or on your unsubscribe list.
- No agent of record is assigned, so nobody owns the approval.
- The same household was asked inside your cooldown window. Duplicates happen when a couple appears as two contacts.
- The client is flagged as a vendor, lender, relative of the agent, or another conflict. Google Business Profile Help lists reviews based on a conflict of interest as rating manipulation.
Texting is its own question. Whether you may text a past client depends on the consent you recorded and on laws we are not qualified to advise on. Ask your attorney, then add a "consent on file" field and block texts without it. Email is the simpler place to start.
Message blocks: the wording is the problem
The check scans the final text, including any edit the approver made, and blocks on these:
- Any offer of a gift, raffle entry, discount, closing credit or gift card.
- A request to name a specific agent, or to include specific content in the review.
- A request for a number of stars.
- Wording that screens the client first, such as "if you were happy with us." Send the same neutral ask to everyone who passes the data checks.
The same Google page says merchants should not pressure people to leave reviews while on the premises. In real estate that matters at the closing table. Keep the ask out of the room and send it afterward.
The hard case: the client with an open complaint
This is where the rules pull against each other. You do not want to email a review request to someone who is angry about a repair credit. But Google's policy bars selectively soliciting positive reviews. Silently dropping unhappy clients from the list looks a lot like that.
We do not have a definitive answer, and Google does not publish how it judges edge cases. Here is the approach we would take, and you should check it against the current policy yourself.
- Do not use a sentiment score, a star guess or an AI judgment of tone to decide who is asked. That is the screening Google's policy describes.
- Use a plain, human-entered flag: "open issue, manager is handling." A person sets it and a person clears it.
- A flagged row does not vanish. It moves to a "call first" lane with the reason showing, and the manager decides when, or whether, it re-enters the queue.
- Record who set the flag and why, so you can show the rule was about an unresolved matter and not about a prediction.
The honest limit: a flag that is applied only to unhappy clients still has the effect of favoring positive reviews. Talk it over with your broker before you rely on it. The safest default is the narrow one. Flag open disputes only, and review the flag list monthly.
Pacing the queue so it does not look like a spike
On day one you will have a backlog: every closing from the past year or two, all eligible at once. Do not release them together.
The Google Business Profile Help page lists unusual volumes or patterns of review contributions as a sign of attempts to manipulate a rating. We do not know what volume Google treats as unusual, and nobody outside Google does. A per-week cap is a cheap way to stay on the safe side. Set it low and raise it slowly.
The cap also keeps the approver's job small. Twenty rows a week gets read. Four hundred rows gets bulk-approved without reading, which defeats the point of the queue. If you find yourself clicking "approve all," lower the cap.
Timing ties to the rest of your follow-up plan. Review requests should not land on the same day as a home anniversary email or a market update. Our January-to-March follow-up calendar for past clients shows how to space those touches ahead of the spring 2027 market. Slot the review ask into a week with no other message.
Where AI fits, and where a person stays in charge
Automation here is mostly plain rules: status, dates, flags, caps. AI is useful in one small place. It can draft the opening line of the message from CRM fields like the street name, the closing month and whether the client was a first-time buyer. The approver reads that line and can change it.
AI does not decide who is eligible, does not send anything, and does not rate how a client feels. If you want to see how we think about approvals and exception handling in general, that is what our AI workflow automation work is built around: a human sees the exception before it leaves the building.
The same personalization fields also feed home anniversary emails, so pulling them once helps both.
The profile link itself is a local search matter. A review request sent to a profile with the wrong category or a stale address helps nobody. Run a Google Business Profile audit before the first batch goes out.
What this does not do
- It does not make clients reply. Some people never will.
- It does not promise any star rating, review count or ranking change.
- It does not replace your broker's judgment on testimonials or on what your state and local association expect. Ask them.
- It does not fix a dirty CRM. It shows you the dirt.
If you want this built into a brokerage CRM, that is the work on our real estate industry page: lead routing, agent follow-up and transactions in one system, with the queue as one screen inside it. It can also be built on top of the CRM you already run. Either way, the approval rules above are the part to decide first.
Frequently Asked Questions
Can real estate review request automation send requests without a person approving them?
Technically yes, but we would not build it that way. A review request is public-facing and uses client data that is often wrong, such as closing status, contact details or an unresolved dispute. A human approval queue catches those errors before they send. Rule-based checks still run first, so the approver only sees rows that pass the basic tests.
Is it allowed to only ask happy clients for Google reviews?
Google's policy says merchants may not selectively solicit positive reviews or discourage negative ones. Sending one neutral request to every eligible client is the safer pattern. Check the current policy page yourself, because Google edits it. This is general information, not legal advice. Ask your broker or attorney about rules specific to your state.
Who should approve review requests at a small brokerage?
The agent of record for their own clients, with the operations lead or transaction coordinator as backup and for stale rows. Template edits and check rules should sit with one or two admins. Log every approval with the name, time and exact message text, so you can answer a client complaint quickly.
Can I offer a closing gift or discount in exchange for a review?
Google's Business Profile policy prohibits offering incentives such as payment, discounts or free goods and services in exchange for a review. A closing gift given with no review attached is a different thing, but keep the two messages separate and never link them. If you are unsure, ask your broker before sending.
How many review requests should I send per week?
We do not have a sourced number, and Google does not publish one. Start low, such as a batch small enough that the approver actually reads each row, and raise it slowly. Release a backlog over weeks, not in one day, so the pattern of new reviews looks like your actual closings.
Want the queue built around the CRM and the team you already have? Talk to us about scoping it, starting with the six fields and the approval roles above.