Review-Response Drafter | Real Minds AI

Where it works well

It drafts a specific, on-brand reply to every review the same day, and flags the ones a manager must handle personally.

Done by hand a genuine reply is several minutes of manager time each, across several platforms — the first thing dropped on a busy service, so reviews pile up unanswered.
Best for an operator carrying steady review volume across multiple platforms — a busy restaurant, café or multi-site group — where prompt, specific replies move both ranking and guest trust.
One published case cut the per-review drafting task from about 30 minutes to 5; the recaptured time goes back into the guest in front of you, not the inbox.

The slow, invisible cost is the reply that never gets written — a venue gets reviews across Google, the delivery apps and the booking sites faster than anyone has time to answer, so the work that protects reputation is the work that keeps getting deferred, and a generic "thanks for your feedback" reads worse than silence.

Where it works badly

It is confidently fluent when it should be cautious — a polished reply to a review it didn't really understand reads worse than no reply.

Weak on the reviews that matter most — allegations of illness, food safety or a named staff member need a manager's judgement and often a private channel, not a public reply. It should flag, not draft a smooth answer.
It can quote a detail back convincingly while missing sarcasm, a mixed signal, or a complaint that is really about something it can't see in the text. A specific-sounding reply isn't the same as an accurate one.
If your reviews are low-volume or each one genuinely needs the owner's personal voice, hand-writing them already does better than a draft you have to heavily rework.

The honest test

If you would not be comfortable with this reply going public exactly as drafted — for the worst review in the queue, not the easy five-star one — then the draft has to wait for a person, and the tool is a head start, not an answer.

Feed it a one-star review alleging someone got sick and it will happily draft a warm, well-worded apology — which is exactly the wrong move when an admission of fault is a legal and food-safety matter, not a tone exercise. Fluent and wrong is the trap.

What it doesn't do — and shouldn't

It drafts the reply and flags the risky ones. A manager reads, edits and posts. That boundary is deliberate.

WHAT IT DOES

Drafts a reply matched to the rating, quoting the detail the reviewer raised

Flags any review mentioning food safety, illness or a staff complaint for the manager

Shows the rating and the quoted detail beside each draft so a person can check the match

WHAT IT WON’T

Post any reply to any platform, or reply on its own schedule

Admit fault, offer compensation, or respond to an illness or food-safety allegation

Decide whether a review is fake or breaches a platform's policy

A public reply is a public statement by the business. Under the Australian Consumer Law a venue can't fabricate or selectively curate reviews, and an on-the-record admission about illness or food safety can have consequences beyond reputation — a food-borne illness complaint is a matter for the state or territory health authority that enforces the Food Standards Code, not for a same-day apology. The accountable person stays on the decision because the consequence lands on them — not the tool.

What your data has to look like

Live review feeds with the rating attached, a written brand voice, and clear rules for what must be escalated.

Typical readiness

across orgs we see, before the first job

A connection to each review platform

Needs shaping

The rating attached to each review

Usual weak point

A written brand voice and reply policy

Needs shaping

Escalation rules for risky mentions

Needs shaping

Contact details for the private channel

Usually ready

The real first job

The brand voice and the escalation rules are usually the weak point — they live in the owner's head, not written down, so every manager replies a little differently and nobody has agreed what must never be answered in public. Getting that into a clear, written policy is usually the real first job — larger and more valuable than the drafting layer on top, because it is what keeps the replies consistent and the risky ones out of the public thread. Once it's written, every reply after that is faster and on-brand by default.

Right fit if…

You carry steady review volume across Google, the delivery apps and the booking sites

Replies are slipping — reviews pile up unanswered because nobody has the time

You have, or are ready to write, a clear brand voice and reply policy

You want every reply to quote the specific detail the guest raised

Walk away if…

Your review volume is low enough to answer personally without strain

Every reply genuinely needs the owner's own voice and can't be templated

You have no agreed rule for what must be escalated rather than answered publicly

You want a tool that posts replies automatically with no one reading them

Open questions

The worried-buyer questions, answered straight

It never posts anything itself, and an allegation about illness, food safety or a named staff member is flagged for the manager rather than drafted into a smooth apology. That’s deliberate: an on-the-record admission about a food-borne illness is a matter for the state or territory health authority that enforces the Food Standards Code, not a same-day public reply. For ordinary reviews it drafts a response and shows the rating and the detail it quoted, so a person checks the match before anything goes public.

It works best where the rating is attached and there’s real text to quote; where the signal is thin — a bare star rating, heavy sarcasm, or a complaint that’s really about something not in the text — messy input is where it earns its review step rather than its automation, and it should flag for a person instead of drafting a confident reply. The honest test: if you can’t tell from the review alone what the guest actually means, neither can this.

No. It removes the blank-page drafting — the several minutes per review of finding the right words across several platforms — so the manager’s time goes to the judgement: whether the tone is right, whether a flagged review needs a private call, whether to offer anything. Every reply is still read and posted by a person. The capacity it frees goes back into the floor and the guest, not into chasing the review inbox.

Current to the live feed — the value is replying promptly, so it reads new reviews as they land rather than working from a stale export. The brand voice and escalation rules behind it also have to be kept current: if your menu, your team or your policy on what to say publicly changes and the rules don’t, the drafts drift out of step. It is not a set-and-forget tool; the policy it drafts against is maintained as part of the build.

Review text, reviewer names and any contact details are handled within the controls you set, not pooled or sent anywhere on the tool’s own initiative, and nothing is published without a named person approving it. A guest’s name and the substance of their complaint carry obligations under the Privacy Act, so we scope where that data sits and who can see it as part of the build. The demo here runs entirely on fabricated data; Sarah M. and Karen L. are not real reviewers.

It can flag a review that reads as suspicious, but it does not decide a review is fake, dispute it, or remove it — and it will never draft fake positive reviews to bury a bad one. Under the Australian Consumer Law a business can’t fabricate or selectively curate its reviews, and challenging a genuine-looking review is a platform process and a judgement call for a person. The tool surfaces; whether and how to contest a review stays with you.

What it takes to build

3–4 weeks · 4 phases

Reused from template~65%

Bespoke to this skin~35%

stack · Claude · review-platform APIs · review UI

What it would cost

Fixed scope, fixed price, fixed dates.

Bite-sized first piece

One contained change, low risk

Pilot build

Most builds land here

Embedded support

Scale on proof

Considering this for your venue?

The honest place to start is a bite-sized first piece — one platform, your own brand voice, low risk. Tell us where it hurts; we'll play it back, scope it, and show you what's possible.

Book a call How we work