AI assistant · N5 Complete system
Automated Comment Reply Builder for Restaurants
A comment isn't answered for the person who wrote it: it's answered for the people reading. This writes the three cases that actually hurt in public — fair criticism, the troll, and the price question in front of everyone — with the second-reply rule, the single-reply rule and the rule that a price gets stated with a figure.
N5Complete system
Several chained pieces you install and operate over time, with their review cadence. It changes how the business runs, not one decision.
◆ This assistant 110 of the 164 published assistants are at this levelIs this for you?
What it does and what you get
- The expensive mistake isn't being slow: it's answering the complainer well and the onlookers badly.
- «DM us» calms one person and tells the other two hundred there is something that can't be said out loud; the witty reply wins the argument and loses the customer reading it.
- This assistant leaves you the written replies for the three cases that jam in public, each with its first line, its second line and the exact moment the thread closes or moves to private.
- The common law — tone, timing and escalation — comes from the sister piece; this one comes down to the channel and writes.
What you end up with
The three-case matrix with the signal that gives each one away, the first reply, the second or the silence, and who signs; the reply bank as three cards with two openers and two closers each and one version per channel; and the thread checklist with what good looks like, what bad looks like, and the log template to review the outcome at 7 days.
EXAMPLE This is what you getAn example built with Sazón de Origen, the house's fictional restaurant.Not your data: your result is built with YOUR restaurant's.See the full example →Sample result
The owner asked this: «We're Sazón de Origen, contemporary Colombian cooking in Medellín, two locations, six years in. On Sunday everything landed at once on Instagram: a lady commented that she waited 40 minutes for a house platter and she is absolutely right; a guy who has never been wrote under the three-meat sancocho video that we charge restaurant prices for ordinary food; and there are eleven comments asking how much the garlic trout costs. I answered the lady with ‹DM us› and it got worse: two more people said it happened to them too. Tell me what to write, in my voice — we're warm and unpretentious.»
| Case | Signal that gives it away | First reply | Second reply or silence | Who signs |
|---|---|---|---|---|
| Fair criticism — 40-minute wait | States a verifiable fact with day and time, and has eaten here before | Public, acknowledges the wait and says what changes on Sunday | Public too, and only then: WhatsApp with a real name and a 24-hour window | Manager, original location |
| Troll — «restaurant prices for ordinary food» | No order history, opines on price without tasting, fishing for a reply | One only, dry, no irony: a fact and an open door | None. The thread is left alone for 48 hours | Community manager |
| Price question — garlic trout, 11 times | «How much is it?» repeated under the same video | Exact figure, portion size and what it includes, pinned at the top | Only if they ask about delivery: the channel price, commission included | Community manager |
Illustrative figures from a fictional case. All three replies are written for the 4,200 people who watched the video, not the 13 who commented.
First reply, public (published today, before the thread sinks):
Doña Marta, forty minutes for a house platter is too long and there's no excuse: on Sunday lunch the delivery orders and the dining room hit us at once, and the kitchen couldn't keep up. From this Sunday the platter goes out with the meats already portioned from 11:30, and we're adding one more person on the pass from 12 to 3. Thank you for saying it here instead of leaving quietly.
Second reply, public, next day (this is where it moves to private, not before):
Doña Marta, we owe you one: Laura will message you on WhatsApp today before 6 p.m. to invite you back to try the change. If anyone else waited that Sunday, write to us and you'll get the same answer.
What was NOT written, and why:
| Discarded line | Why it doesn't go public |
|---|---|
| «We regret any inconvenience caused» | Says neither what happened nor what changes; the reader concludes nobody took responsibility |
| «DM us and we'll sort it out» as the first move | That is exactly what was done on Sunday, and two more people replied «same here»: in the open it reads as hiding |
| «We had a very busy day» | Turns the customer's problem into the house's achievement |
Illustrative example generated with a fictional restaurant. Not a promise of results, not a client case.
This is 2 of 3 parts. Behind the wall sit **the troll reply, the price answer and the thread log**: the single line written to someone looking for a fight, the closing line in case they push, the price with portion size and the delivery-app figure stated separately, and the table where Thursday shows what happened at 7 days with each of the three cases.
See the full example — free accountWorks with these AIs
How to use it
Which of your data it uses
This assistant works with 13 of the 23 fields in «My Restaurant», among them Restaurant name, Business type and City and country. It does not ask out of curiosity: these are what make the answer speak about YOUR scale instead of an industry average — the same calculation on a business of another size returns a number you cannot decide with. You fill them once and they apply across the library, so the second assistant you open already starts with them in place. Whatever you leave blank, the assistant asks for one thing at a time instead of inventing it, and says so before answering, so you know what it is working from.
Always up to date
Every assistant carries a visible version and date. When an AI changes how it works, the assistant's version goes up and you see what changed.
v1.0.12026-08-27It now declares the «See photos» capability, so the AI selector warns before you open the prompt in one that lacks it, instead of letting you find out mid-task.v1.0.02026-08-20Initial version. Born in the August 2026 marketing batch for the channel where everything is read over someone's shoulder: the public comment. The piece was written around three rules that make people uncomfortable and are therefore rarely followed — a fair complaint moves to private on the second reply and never the first, a troll gets one answer and never two, and a price is stated with a figure instead of dodged with an «I'll DM you» that, in front of witnesses, sounds like there is something to hide.Same method, another variant
Inside the app this is a single control: you press it and the assistant reframes itself, including the variants that don't appear here because they don't have their own page yet.
More assistants in this category
The job description that gets signed, not the task list that gets forgotten: what belongs to the community man…
N5 · Complete systemThe board that rules on whether the social media role is paying for itself: every indicator with a written for…
N5 · Complete systemChoose at the top where the message lands: a direct message or a public comment. In this one —the direct messa…
Included in the always-growing library
Access to every published assistant, adapted to your AI and personalised with your restaurant's data, updates included.
