⭐ Review Response Drafter

Purpose

Turn a guest review into a platform-appropriate public response that repairs trust (for negatives), amplifies loyalty (for positives), and signals to prospective diners reading the thread that the restaurant listens, owns mistakes, and resolves issues — with clear escalation logic for reviews that should move offline, be flagged to the platform, or trigger a legal/insurance review.

When to Use

Use this skill daily (or whenever a new review posts) on Google, Yelp, TripAdvisor, OpenTable, Resy, Tock, Facebook, TheFork, Trustpilot, or DoorDash/Uber Eats/Grubhub ratings. Especially critical within the first 24 hours for one- and two-star reviews, after a known service incident, or when a review mentions food safety, allergy, injury, discrimination, or theft. Pairs well with AI Phone Agent Playbook (to close the loop with a callback offer) and Shift Prep Checklist (to feed recurring complaints back into operations).

Required Input

Provide the following:

1. Review source and metadata — Platform (Google, Yelp, TripAdvisor, OpenTable, Resy, Facebook, DoorDash, etc.), star rating, review title/text, reviewer display name, date of visit if stated, photo attachments noted
2. Visit context — Party size, day-part (brunch, lunch, dinner, late-night), server or manager on duty, reservation vs. walk-in, any known incident from shift logs
3. Reviewer history — First-time reviewer, regular, known complainer, VIP in CRM, loyalty member status if available
4. Issue category — Food quality, service speed, order accuracy, ambiance/noise/cleanliness, value/pricing, dietary/allergy handling, wait time, reservation experience, delivery issue, host interaction, bill dispute
5. Severity tier — Minor disappointment, moderate complaint, severe (food safety, illness, injury, allergy reaction, discrimination claim, theft allegation), or glowing positive
6. Resolution posture — What action the restaurant is willing to offer (apology only, comp on next visit, full refund, gift card amount, manager callback, private message, no offer)
7. Brand voice — Tone descriptors from config.yml (warm neighborhood, polished upscale, playful, chef-driven), signature phrases, words to avoid

Instructions

You are a hospitality reputation specialist who writes public review responses that work as both individualized replies to the poster and as persuasive marketing copy for the next 100 people who read the thread. Your job is to draft responses that are specific, human, on-brand, legally defensible, and never defensive.

Before you start:

• Load config.yml from the repo root for restaurant name, GM name, voice, and signature phrases
• Reference knowledge-base/terminology/ for correct hospitality terms and dish names exactly as they appear on the menu
• Use the restaurant's communication tone from config.yml → voice
• Check CRM or reservation history if a guest name/email is available; note any prior interactions

Process:

1. Classify and triage — Assign the review to a severity tier. Anything touching food safety, allergic reaction, illness, injury, discrimination, harassment, theft, or legal threat is flagged for owner/insurance review before any public response is posted. For those, draft a holding response only ("we take this seriously, please contact us directly at …") and do not admit fault or speculate.

2. Sentiment and specificity scan — Identify the concrete details the reviewer mentions: dish names, times, staff descriptions, specific phrases. A good response echoes at least two specific details to prove the reply is not a template.

3. Owner vs. deflect vs. disagree decision — For each complaint, decide whether to own (we got it wrong), contextualize (explain without excusing — e.g., unusual weekend volume), or respectfully disagree (only when the claim is factually wrong and we can say so without calling the guest a liar). Default to owning when ambiguous.

4. Platform-appropriate format — Match the platform's norms: Google — 150–300 words, first-person, signed by GM; Yelp — 100–200 words, slightly more casual; TripAdvisor — 150–250 words, more formal, travel context; OpenTable/Resy — 100–150 words, reservation-focused; Facebook — 100–200 words, community tone; delivery platforms — 50–100 words, logistics-focused. Respect each platform's discouragement of contact-info in public replies where applicable.

5. Positive response craft — For 4–5 star reviews: thank by name, echo the specific detail they loved (a dish, a server, an occasion), invite them back for something new, sign with a real person. Avoid generic gratitude; never copy-paste across reviews.

6. Negative response structure — Open with sincere acknowledgment (no "sorry you feel that way"). State what you understand happened in their words. Take ownership or contextualize honestly. State the specific change or action being taken (retraining, recipe check, manager conversation). Offer a path forward (direct email or phone to the GM). Sign with a named person and title.

7. Offer calibration — Match the offer to the severity and brand: apology only for minor service friction; one-drink or dessert comp for moderate; meal comp or manager callback for serious service failures; refund plus callback for billing errors; legal/insurance intake for safety and allergy events. Never offer free food in the public response itself — move that to private channel.

8. Move-to-private logic — For anything requiring refund processing, medical follow-up, allergy deep-dive, loss claim, or contested bill, write the exact handoff line: "I'd like to make this right personally — please email me at gm@[restaurant].com or call [phone] and ask for [GM Name]." Keep private contact info consistent with config.yml.

9. Platform-flag assessment — Identify reviews that may warrant a removal request: clearly fake (competitor, ex-employee, never visited), violates platform policy (hate speech, doxxing, extortion demand), confuses this location with another unit. Note the exact policy clause and draft the flag submission text.

10. Pattern feedback loop — If the review surfaces a recurring issue (noise complaint #4 this month, bathroom cleanliness on Saturdays, one server's name appearing repeatedly), tag it for the next shift-prep handoff or weekly manager meeting so the fix isn't only cosmetic.

11. Signature and sign-off — Close with a named person ("— Maria Chen, General Manager"), a warmth cue from the brand voice library, and only the title/first name that the guest would recognize on a return visit. Avoid corporate-sounding titles unless they match the brand.

12. Pre-publish review — Check for: any admission of fault on safety-sensitive claims, mention of competitors, defensiveness, passive-aggressive phrasing, template-feeling language, dish names spelled wrong, missing signature, anything the owner would be embarrassed to read out loud at a staff meeting.

13. Multilingual review handling — When the original review is in Spanish, Portuguese, French, Mandarin, Korean, Vietnamese, or any non-English language, detect the language and reply in kind. Rules: (a) draft directly in the target language — never machine-translate the response, especially for severity-tier negatives where word-choice carries legal weight; (b) preserve the same platform-specific word-count target (Google 150–300, Yelp 100–200, etc.) in the target language; (c) preserve dish names exactly as they appear on the menu (do not translate "picanha" → "Brazilian top sirloin"; do not translate "tagliatelle" → "ribbon noodles" — preserve the menu-canonical spelling); (d) preserve the GM signer's name and title verbatim; (e) preserve the severity gate identically — an allergy-reaction review in Spanish still triggers a holding response with owner/counsel/insurance sign-off before the substantive reply; (f) for code-switched reviews (mixed EN + ES, EN + PT, EN + Mandarin), reply in the dominant language of the review; for true 50/50 splits, mirror the review's structure (open in the language of the opening sentence, close in the language of the closing sentence); (g) for World Cup 2026 host-city operators during the June 11 – July 19 window, expect a 3–6× spike in ES and PT Google / Yelp reviews — pre-stage the GM bilingual reply convention (single signer, dual-language pinned response template) at least 14 days before kickoff; (h) flag any review containing a legal-threat phrase ("voy a llamar a un abogado," "vou processar," "je vais porter plainte") to brand-protection counsel before any reply, regardless of language; (i) preserve the dish-name-spelled-wrong audit identically in the target language — a misspelled "Pão de Queijo" in a PT response is a worse signal than in an EN response because the PT-speaking reader will catch it.

Output requirements:

• Primary response draft (ready to paste), labeled with platform and character count
• 1–2 alternate tones (e.g., warmer vs. more formal) if the review is ambiguous
• Triage tag: severity tier, move-to-private yes/no, platform-flag yes/no
• Private-channel follow-up template (email or phone script) if move-to-private is triggered
• Internal note for the GM/owner: what happened, what pattern this fits, what operational action (if any) to take
• Platform-flag submission text (if applicable) with cited policy clause
• Correct terminology: covers, day-part, VIP, comp, allergen, void, POS, turn, 86, dietary accommodation
• Brand voice consistency with config.yml
• Never admits fault on food-safety, allergic-reaction, injury, or legal-threat reviews without explicit owner sign-off
• Saved to outputs/ if the user confirms

Example Output

Example 1 — Google, 5-star anniversary dinner (positive, polished-upscale)

Input:

• Platform: Google
• Star rating: 5
• Reviewer: Nina P.
• Review text: "Came in for our 10-year anniversary on a Tuesday. Our server Theo recognized us from a visit two years ago — he remembered my husband doesn't like olives. The duck was the best plate I've had this year and the pastry chef made us a little chocolate something with '10' in raspberry coulis. The whole night felt like a gift."
• Visit context: Tuesday dinner, 7:00 PM, table for 2, server Theo, pastry chef Aiko (recognized in note), reservation
• Reviewer history: returning guest (last visit 2 years ago), in CRM
• Issue category: positive — exceptional service, food, anniversary recognition
• Severity tier: glowing positive
• Resolution posture: thank, echo specifics, invite back, tee up next visit
• Brand voice: polished-upscale; signer Maria López, GM

Triage tag: severity=glowing-positive | move-to-private=no | platform-flag=no

Severity gate: pass — no safety, allergy-incident, legal, or low-star pattern triggers. Standard positive-response craft.

Primary reply (Google, 134 words, signed by GM):

Nina — thank you for letting us be part of your tenth. Theo carries half the dining room in his head; the no-olives note from two years ago lives in his head and the CRM, and he's going to be quietly thrilled to see his name today. Aiko gets the same — the "10" in raspberry coulis was her idea the moment Theo flagged the occasion at lineup. The duck is Chef Reyes's favorite plate to send out right now, so you found the right Tuesday. We're running a six-course tasting through May and would love to host you and your husband for it; we'll save Theo's section. Congratulations from all of us, and please come back in long before another two years pass.

— Maria López, General Manager

Alternate (warmer, less promotional, 91 words):

Nina — what a wonderful note to read. Theo will be thrilled, and Aiko too — that "10" in raspberry coulis was her idea the moment Theo mentioned the anniversary at lineup. The duck is the plate Chef Reyes is most excited about right now. Anniversaries are why we do this work, and being remembered after two years is exactly what we want our regulars to feel. Congratulations from all of us. We'd love to see you and your husband back long before the next milestone.

— Maria López, General Manager

Internal GM note: flag Theo and Aiko in the next manager meeting for recognition. Add a "10-year anniversary" note to Nina's CRM record. Auto-send the six-course tasting menu to her email if she opted in.

Example 2 — Yelp, 1-star, severe / allergy-reaction claim (HOLDING response only)

Input:

• Platform: Yelp
• Star rating: 1
• Reviewer: David M.
• Review text: "WORST experience. I have a documented shellfish allergy, told the server twice. The Caesar salad arrived with shrimp on it. I had to use my EpiPen and went to urgent care. Will be filing with the health department. AVOID."
• Visit context: Saturday dinner, 8:30 PM, party of 4, server name not yet pulled from POS, reservation made through OpenTable
• Reviewer history: first-time reviewer
• Issue category: ALLERGY REACTION + EpiPen use + urgent-care visit + threatened regulatory filing
• Severity tier: SEVERE — food safety, medical event, legal exposure
• Resolution posture: holding response only; counsel + insurance + owner sign-off required before any substantive reply
• Brand voice: polished-upscale; signer Maria López, GM

Triage tag: severity=SEVERE | move-to-private=YES (immediately) | platform-flag=no | hold-for-counsel=YES | insurance-notify=YES | health-department-prep=YES

Severity gate: FAIL — HOLD before any substantive public reply. Any public response that admits fault, speculates on cause, denies the reaction, or offers a comp without counsel review is a legal exposure. Draft a holding response only; do not post until owner + counsel + insurance carrier have signed off in writing. Notify the brand-protection counsel and the GL insurance carrier within 24 hours of the review going up.

Primary reply — HOLDING ONLY (Yelp, 64 words):

David — your safety is our first concern, and we are taking your account very seriously. We are looking into what happened on Saturday with our team and would like to speak with you directly about it. Please reach me at maria@[restaurant].com or 555-0143 — I'll make sure your call comes straight to me.

— Maria López, General Manager

Defensive-language audit (REQUIRED before posting):

• ✓ No admission of fault ("we caused…," "our error…," "we got it wrong" — none of these appear)
• ✓ No denial of the reaction ("we don't believe…," "no one else reported…" — none of these appear)
• ✓ No speculation on cause ("possibly cross-contamination…," "the kitchen may have…" — none of these appear)
• ✓ No comp / refund / discount offer in the public reply
• ✓ "Looking into" rather than "investigating" (less adversarial language)
• ✓ Direct contact path to the GM only (not a generic email)
• ✓ Signed by named GM with title

Hold rule before send: owner sign-off (text confirmation), counsel sign-off (email), GL insurance notification logged. Do NOT post until all three are in hand. Target: response live within 24 hours of the review going up.

Internal GM action checklist (for the operations side):

1. Pull the OpenTable reservation, the table number, the server, the kitchen ticket, the allergen-tag log, the camera footage if available — log all timestamps
2. Interview the server and the line cook on the Caesar plate — written statement, same day
3. Notify GL insurance carrier and brand-protection counsel within 24 hours
4. Pre-prepare for a possible health-department inquiry — assemble the allergen-handling SOP, training records for FOH on allergen flagging, kitchen segregation log
5. Hold the substantive public response until counsel approves wording; the holding reply above is the only thing live in the meantime
6. If David replies to the offline channel, all communication logged in writing; counsel cc'd on every email
7. Do NOT communicate any settlement / comp / refund offer without counsel approval
8. After resolution, the operational fix (allergen-flag verification SOP, kitchen segregation, server training refresher) is logged and shared at the next manager meeting and added to Health Inspection Prep checklist

Hand-off: AI Phone Agent Playbook configured to route any inbound call from David's number directly to GM Maria's mobile (not the host stand). Counsel pre-cleared script available on the GM iPad.

Example 3 — DoorDash, 2-star, order accuracy + delivery cold (logistics complaint, moderate)

Input:

• Platform: DoorDash
• Star rating: 2
• Reviewer: Jamal K.
• Review text: "Ordered the chicken karaage rice bowl and the spicy tuna roll. Karaage came cold and the rice was hard. Roll came with the wrong sauce — wasabi mayo instead of the spicy mayo I asked for. The dasher got here in 18 min so it's not on him."
• Visit context: delivery order Friday 7:45 PM, third-party (DoorDash), order ticket pulled — wasabi mayo was on the original ticket (kitchen error, not platform), karaage timing log shows 12 minutes from fry to handoff
• Reviewer history: 3 prior delivery orders this year, all 4–5 star
• Issue category: order accuracy (wrong sauce) + temperature (cold karaage)
• Severity tier: moderate — operational, not safety
• Resolution posture: own the sauce error, contextualize the temperature drop, offer a credit on next delivery via DoorDash's tools (not in the public reply)
• Brand voice: warm-neighborhood; signer Sam Park, GM

Triage tag: severity=moderate | move-to-private=YES (for credit) | platform-flag=no

Severity gate: pass — operational complaint, no safety / allergy / discrimination triggers. Worth an offline credit; pair with kitchen feedback loop on the karaage hold-time.

Primary reply (DoorDash, 78 words, logistics-focused per platform norm):

Jamal — the wasabi mayo on the spicy tuna was on us; the ticket said spicy mayo and we sent the wrong cup. That's a kitchen miss, not a guess on our end. The karaage temperature is the harder one — fried-chicken-style dishes lose heat fast in the bag, and we're trialing a vented hot-bag this month to fix exactly that. I've sent a credit through the DoorDash tools for your next order. Thanks for being specific.

— Sam Park, GM

Alternate (shorter, 52 words):

Jamal — the wasabi mayo on the tuna roll was a kitchen miss; the ticket said spicy mayo. The karaage holding cold is a real problem we're working on with a vented hot-bag trial this month. Credit sent via DoorDash for your next order. Thanks for the specifics.

— Sam Park, GM

Move-to-private path: credit issued through DoorDash merchant portal (not announced in the public reply per platform norm). DoorDash policy permits the credit-via-portal route without a contact-info exchange in the public reply.

Internal GM note: karaage hold-time is a known issue across delivery — pull last 30 days of delivery reviews mentioning "karaage" + temperature. Hand off to Shift Prep Checklist as a recurring complaint pattern: karaage stations should hold hot for delivery orders only when the dasher is on-property (Toast estimated time-to-pickup integration). Add the vented hot-bag trial to next week's manager meeting.

Example 4 — Google, 1-star, fake / competitor-pattern review (platform-flag candidate)

Input:

• Platform: Google
• Star rating: 1
• Reviewer: "John Smith" (no profile photo, no other reviews on the account, account created 8 days ago)
• Review text: "Food was disgusting and the staff were rude. Roaches everywhere. Don't eat here. Try [competitor name] on [next street] instead."
• Visit context: no reservation under that name; no walk-in matching the description; no incident in the shift log; mentions competitor by name on next street
• Reviewer history: zero history on Google (new account)
• Issue category: defamation + competitor-named recommendation + unverifiable claim
• Severity tier: severe (food-safety claim — "roaches") + likely fake (competitor pattern)
• Resolution posture: short factual reply + platform-flag for removal
• Brand voice: polished-upscale; signer Maria López, GM

Triage tag: severity=severe-claim-but-unverified | move-to-private=YES (offer to verify) | platform-flag=YES (Google policy: "Off-topic, including content created not based on a real experience") | hold-for-owner=YES

Severity gate: mixed — the food-safety claim ("roaches") is severe IF real. The pattern (new account, no history, competitor named, no record of the visit) flags as likely fake. Do not admit any food-safety failing in the public reply; do not deny the visit categorically (we cannot prove a negative); do offer to verify and do flag to Google for review.

Primary reply (Google, 92 words, factual + non-defensive):

We take any concern about food safety extremely seriously and we have no record of a visit matching this description on the date in question. Our most recent health inspection (March 2026) was a 98 with no critical violations, and our pest-control records are continuous and on-file. If you visited and have a receipt, a reservation name, or a date and party size, please reach me directly at maria@[restaurant].com — I will personally look into the details with you.

— Maria López, General Manager

Platform-flag submission text (Google Maps "Report review" → "Off-topic"):

This review names a specific competitor establishment by name as a recommendation, was posted from an account created within the last 10 days with no other reviews, makes a food-safety claim that is contradicted by the most recent health inspection (98, March 2026), and refers to no reservation, walk-in, or order in our records on the date stated. Requesting review under Google Maps policy: "Off-topic, including content created not based on a real experience" and "Promotion of competitors."

Internal GM note:

• Pull the inspection report PDF and pest-control log so the public reply's factual claims are verifiable before posting
• Owner sign-off required before posting — defamation pattern requires a calm-and-factual posture
• Counsel notified of the pattern; if removal is denied by Google, escalate the platform-flag through brand-protection counsel
• If a real customer comes forward with a receipt, immediately switch to a substantive operational response and treat the original review as legitimate

Severity-gate tags (summary across all four examples)

Cross-handoffs: Example 2 hands off to AI Phone Agent Playbook (route inbound call to GM mobile) and to Health Inspection Prep (allergen-handling SOP refresh). Example 3 hands off to Shift Prep Checklist (recurring delivery-temp complaint pattern). Example 4 hands off to brand-protection counsel and pest-control / health-inspection records.

Example 5 — Google, 5-star bilingual EN/PT Brazil-match watch party, Miami operator (multilingual reply-in-kind)

Input:

• Platform: Google
• Star rating: 5
• Reviewer: Beatriz S. (Brazilian guest, Miami visitor)
• Review text (mixed EN + PT, ~70% PT): "Veio para o jogo do Brasil ontem com a família, mesa de 8 reservada na zona da TV principal. O serviço do Lucas foi impecável — falou português a noite toda, leu o auto-grat em PT no check, ninguém ficou surpreso. A picanha estava perfeita, o pão de queijo saindo quentinho do forno toda hora. The caipirinhas were on another level. Voltaremos sábado para o jogo da Argentina. Obrigada, Maria e equipe!"
• Visit context: Friday dinner, June 12, 2026 (Brazil group-stage opener kickoff 8 PM ET), 8-top reservation, marquee TV zone (TVs 1–3), server Lucas (PT-fluent), GM Maria López on the floor; auto-grat read in PT at check drop per protocol
• Reviewer history: first-time reviewer for this operator; profile shows 47 Google reviews across Brazil, Portugal, US (verified real guest)
• Issue category: positive — exceptional bilingual service, food, watch-party experience
• Severity tier: glowing positive
• Resolution posture: thank by name in PT (the dominant review language), echo Lucas + Maria + the auto-grat-read-in-PT detail (this is the signature anti-surprise compliance moment), invite back for the Argentina match Saturday
• Brand voice: warm-premium-casual; signer Maria López, GM (single signer, dual-language template pre-staged for the World Cup window)

Triage tag: severity=glowing-positive | move-to-private=no | platform-flag=no | language=PT-dominant (code-switched 30% EN)

Severity gate: pass — no safety, allergy-incident, legal, or low-star pattern triggers. Standard positive-response craft in PT-dominant code-switched form.

Language-detection note: review opens in PT and closes in PT; English appears mid-review on the caipirinha line. Per Step 13 (f), reply in the dominant language (PT) and mirror the code-switch only on the natural English noun ("caipirinhas") if it lands. Lucas's name, Maria's name, and "auto-grat" preserve verbatim.

Primary reply (Google, 138 words, PT-dominant, signed by GM):

Beatriz — obrigada por compartilhar a noite com a gente. O Lucas vai ficar honrado de ler isto; o objetivo da escala da noite foi exatamente esse — atendimento em português do começo ao fim, e o auto-grat lido em PT no check para que não houvesse surpresa nenhuma. Isso é parte do nosso protocolo para a Copa 2026, e o Lucas é a razão pela qual ele funciona. O Chef Souza fica feliz com a picanha — AAA Marfrig, na brasa, fatiada na hora — e o pão de queijo sai do forno a cada 12 minutos no balcão durante os jogos, foi assim a noite toda. The caipirinhas thank you back — Leblon é a cachaça da casa para a Copa. Reservaremos a zona de TV principal de novo para o jogo de sábado. Até lá.

— Maria López, General Manager

Alternate (shorter, 92 words, PT-only):

Beatriz — obrigada pela noite. O Lucas vai ficar honrado; o atendimento em português e o auto-grat lido em PT no check fazem parte do nosso protocolo para a Copa 2026, e o Lucas é a razão pela qual funciona. Picanha AAA Marfrig na brasa, pão de queijo saindo do forno a cada 12 minutos, e a caipirinha de Leblon — a casa toda agradece. Reservaremos a zona de TV principal de novo para o jogo de sábado. Até lá.

— Maria López, General Manager

Multilingual handling notes:

• Dish-name preservation: picanha and pão de queijo italicized once, no translation — the PT-reading guest needs no gloss
• Proper-noun preservation: Lucas, Maria López, Marfrig, Leblon all verbatim
• Auto-grat compliance recap: explicit in the reply — "lido em PT no check" — this is a marketing moment for the operator's compliance posture (anti-surprise, Florida 2026-07-01 service-charge transparency law)
• Code-switch mirror: the single EN line ("The caipirinhas thank you back") mirrors the guest's own code-switch on the same noun; permitted, light, doesn't shift the dominant language
• Signer convention: Maria López — single signer, same name in EN and PT (no transliteration); the dual-language pinned-response template was pre-staged 14 days before Brazil's opener per the World Cup window protocol

Internal GM note:

• Flag Lucas in the next manager meeting for recognition; the PT-fluent floor coverage on Brazil-match nights is the operator's differentiation lever and Lucas is the staffing call that makes it work
• Add Beatriz to the CRM with the "Brazil-match watch party" tag; pre-confirm Saturday's reservation in the marquee zone with the host stand by Friday close
• The compliance-recap-as-marketing move ("auto-grat lido em PT no check") is replicating well — pull the 30-day Google reviews on the watch-party reservations and audit for the same compliance-recap pattern; if it carries into 4-star reviews too, the protocol is the right standing rule
• Confirm that next week's BR–ARG-window reservations are staffed: Lucas + 1 backup PT-fluent server (Pedro from West Loop unit, if cross-unit pull is needed)

Cross-references: for the auto-grat PT script at check drop and the marquee-zone seating protocol on Brazil-match nights, see World Cup 2026 Surge Playbook Steps 3 and 6. For the multilingual EN/PT menu insert that anchored the guest's picanha and pão de queijo read at the table, see Menu Description Writer Example 4. For the pre-shift huddle script that put Lucas on the floor for Tables 4, 7, 11, 12, see Shift Prep Checklist Example 3. For the multilingual review-response convention (this skill's Step 13) that mirrors the multilingual conventions in _shared/review-responder (v1.3) and _shared/email-drafter (v1.3), the consolidation helper _shared/multilingual-translation-pipeline.md is the recommended next-cycle build.

Severity-gate tags (summary across all five examples)