- Home
- Skills
- Marketing & SEO
- Review Response: Draft Replies That Land
More screenshots
Works with the AI Agents you already use
Review Response: Draft Replies That Land
Draft review replies a person would sign: triage first, shape by star rating, tune to the platform, de-escalate by rule, and run batches that never read as one template pasted twenty times.
$5
Review Response: Draft Replies That Land
Example session with this skill installed
We run a small bakery chain. Three reviews landed this week: a 1-star saying the sourdough was stale and the counter staff rude, a 3-star saying the coffee is great but weekend queues are brutal, and a 5-star praising a custom birthday cake. Draft all three replies.
- Read your context and instructions
- Compiled the review response
Triage first. The 1-star names two separable problems (product, staff) — annoyance plus a service complaint, not escalation-class. The 3-star is a mixed operational note. The 5-star is specific praise. All three are Google reviews, so concise.
Reply — 1 star (stale sourdough, rude staff): "You're right to call this out — sourdough that sits past its morning bake is on us, and we've tightened the discard time to 2pm so the afternoon loaves are pulled earlier. The counter experience matters just as much; our manager has picked this up with the team on that shift. If you'll give us another try, mention this review at the till and the replacement loaf is on us — and if you'd rather talk it through, our contact is on this listing. Sorry we got both halves wrong." Rubric: specificity 2, honesty 2 (a dated, named change), tone 2, door 2, length 1 — 9/10.
Reply — 3 stars (great coffee, brutal weekend queues): "Thanks for the honest split — glad the coffee lands. The weekend queue is real: Saturday between 10 and 12 is our crush, and from this month a second till opens for those hours. Ordering ahead for pickup skips the line entirely in the meantime." The concern answered with a fact and a dated change, the strength restated in their words, no defense anywhere.
Reply — 5 stars (birthday cake): "This made the kitchen's week — the baker who built your cake remembered the brief word for word. Glad it landed on the day. If there's a next celebration, ask for the same team and we'll pick up right where we left off." Specific praise echoed, one human detail, one light next step, no feature list.
Batch note: three different skeletons (fix-led, fact-led, warmth-led), three different openings, zero shared sentences — a reader scrolling all three sees a bakery, not a template.
Connects securely to your tools. The creator never sees your data.
About this skill
A review reply is read twice — once by the reviewer, once by every future customer deciding whether to trust you. Most replies fail one of those readers: they grovel, they defend, or they read like the same template with the name swapped. This skill writes replies that survive both readings.
It starts with a five-question triage — what exactly are they saying, how hot is it, what do you actually know, where will the reply live, single or batch — and an escalation decision table that routes safety, fraud, legal, and extortion reviews to a holding reply and a human, never to an improvised draft. Then it shapes the reply to the rating: 1–2 stars get the specific problem named, one honest line, a concrete action with a version or a date, and one private door; 3 stars get their concern answered with a fact; 4–5 stars get the specific praise echoed back in the reviewer's own words.
Platform conventions are built in — concise and indexed on Google, conversational on Yelp, professional on G2 and Capterra, short on app stores, guest-first in hospitality. De-escalation runs by rule, not by mood: no defending, no blame shift, no promises without owners, no private details in public. A five-dimension scoring rubric (specificity, honesty, tone, door, length) gates every draft at 7+/10 with no zero rows. Batch mode gets its own machinery — rotating openings, varied skeletons, fresh middle paragraphs, one unique detail per reply — plus the reply's second life: updating the thread when the fix ships, and knowing which reviews to leave alone.
The worked example takes a real 2-star complaint through triage to a scored, shippable reply, and shows how its five sibling reviews that week varied without drifting.
- It drafts; it does not post. Getting the reply onto the platform is still your step.
- Honest replies need honest inputs — if no fix exists or is scheduled, the skill drafts the shorter honest status reply rather than inventing one.
- Safety, fraud, and legal-threat reviews are deliberately routed to a holding reply plus escalation, not a full drafted response.
- It cannot tell you whether a review is genuinely fake; it gives the neutral-reply and platform-report path for suspected cases.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before any payment.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 24 days ago
Permissions
No special permissions declared or detected