AI assistant · N3 Analysis with a verdict
Reservation Campaign Builder for Restaurants
Builds campaigns that fill specific slots in your reservation book: a weekly demand map, a booking motive per slot, announcement and confirmation copy, and a no-show mechanic.
N3Analysis with a verdict
Crosses YOUR numbers and takes a stand, with a table, units and currency. Not options: a recommendation you can defend.
◆ This assistant 24 of the 164 published assistants are at this levelIs this for you?
What it does and what you get
- The 8 o'clock table that went unsold tonight cannot be sold tomorrow.
- This assistant treats your reservations as perishable inventory: it crosses your slow moments with your city's calendar, picks the slots worth the effort and builds the campaign that sells that table, that day, at that hour, with the no-show policy built in from the start.
What you end up with
A classified slot map, a ready campaign sheet per chosen slot with its copy and no-show policy, and a board with three reservation indicators.
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: "Our dining room sits half empty on weeknights and on Saturday middays. Bookings come in by WhatsApp and by phone, we have no platform and no proper book, and the main location has 22 tables."
Your house has no reservation platform and this map does not need one. The occupancy below came from counting closed checks by hour over the last 8 weeks at the main location —22 tables, a 2-hour turn— and not from the feeling last Friday left behind. That is the most common trap: a slot looks slow only because nobody writes down what arrives by WhatsApp.
| Day and hour | Tables available | Occupancy | Status | Potential |
|---|---|---|---|---|
| Tuesday to Friday, 12:00–14:30 | 22 | 87% | Fills on its own | Out of the campaign |
| Sunday, 12:00–15:00 | 22 | 93% | Fills on its own | Out of the campaign |
| Thursday, 19:00–21:30 | 22 | 41% | Sellable | High |
| Saturday, 12:00–15:00 | 22 | 44% | Sellable | High |
| Friday, 19:00–21:30 | 22 | 63% | Sellable | Medium |
| Saturday, 19:00–21:30 | 22 | 76% | Sellable | Low |
| Tuesday and Wednesday, 19:00–21:30 | 22 | 26% | Hard | Discarded |
| Sunday, 19:00–21:00 | 22 | 19% | Hard | Discarded |
Two slots go in: Thursday night and Saturday midday. Friday and Saturday nights are ruled out because they climb on their own and a campaign would only claim tables that were coming anyway. Tuesday and Wednesday nights leave with their number in plain sight: at 26% that is six tables in the whole room, and no announcement fixes it. Tuesday-to-Friday lunch and Sunday midday stay untouched.
The ceiling comes from the floor. Of the 18 operations people, 9 hold this location and 6 work the night: two servers, a host who also runs the till, and three in the kitchen. Each server covers five tables without crossing tickets, so 10 tables out of 22 are committed to bookings and the other 12 stay open for walk-ins. Occupancy and the shift split are assumed until you check them against the tickets and the real payroll. Illustrative figures, built from the data you gave.
The grid does not sort by taste: it crosses what the slot costs to run against what that slot can bill on top.
| High expected revenue | Low expected revenue | |
|---|---|---|
| Light effort (the night shift handles it) | Onto the fixed calendar. Thursday 19:00–21:30 with the house pot for two: a single preparation, 10 committed tables, and the announcement rides on the lunch check and on the own-delivery packaging. It is the only one that repeats without asking permission. | Do it, but do not call it a campaign. A Friday-night message to whoever has booked before. It costs ten minutes and adds two tables. It earns no plan and no separate measurement. |
| Heavy effort (asks for hands and setup) | It goes in, but one slot at a time. Saturday 12:00–15:00, a long table of 6 or more booked 72 hours ahead: tables are joined, the house dessert comes out (1,900 COP per person, 11,400 per table of six against 288,000 COP of sales) and what you hand over is access, not a discount. It starts once Thursday has three measured weeks behind it. | Say no. Tuesday and Wednesday nights with separate setup. At 26% that is six tables; the night shift starts paying for itself near fourteen, and the eight missing ones are not out on the street on a Tuesday. Leave it as it is. |
The judgement behind the grid, declared and arguable: one slot at a time, measured three weeks before the next one opens. Filling Saturday night is nobody's achievement; the achievement sits on Thursday at seven. The concession: today the person greeting at the door is the same one running the till, and with a dedicated host at night the bottom-left quadrant would ask for far less, so Saturday could start alongside Thursday instead of waiting three weeks.
Motive: house pot for two, Thursday at 7:00 p.m. It comes from one of your declared reasons for visiting —the craving for home cooking—, not from a catalogue: whoever has lunch here midweek does not want to cook on Thursday.
The finding first. Sancocho de tres carnes, as it sits on the menu, does not fit an offer: at 42,000 COP with a 15,100 cost it runs at 35.9%, above the 34% menu average and well past the 32% ceiling here. The problem there is not the offer, it is the menu price. Since the goal is to raise margin without touching menu prices, the pot was built from cost towards price.
| Item | Cost | Price | Food cost |
|---|---|---|---|
| Sancocho de tres carnes, menu portion | 15,100 COP | 42,000 COP | 35.9% |
| Pot for two, one single preparation | 25,400 COP | — | — |
| Rice, avocado and plantain for two | 2,800 COP | — | — |
| Two panela lemonades | 2,800 COP | — | — |
| House pot for two, complete | 31,000 COP | 98,000 COP | 31.6% |
That is 49,000 COP per person: inside the menu range (28,000 to 72,000) and above the 48,000 average check. It leaves 33,500 COP of contribution per cover against the 31,680 of a normal visit at 34%. Costing assumed until you check it on the scale. Illustrative figures.
Announcement, for the lunch check and the own-delivery insert:
Thursday, 7:00 p.m.: the sancocho de tres carnes comes to the table in the pot, for two, with rice, avocado, plantain and panela lemonade. 98,000 the pot. Ten tables only, held through the same WhatsApp you order delivery from, or by phone. Tonight you are not cooking.
Confirmation, Wednesday 7:00 p.m.:
Good evening, [name]. Tomorrow, Thursday at 7:00, we have your table for two with the house pot. Reply "coming" or "can't make it": either one helps us.
Reminder, Thursday 4:00 p.m.:
[name], we are expecting you today at 7:00. The table is held until 7:20; if something came up, tell us and we pass it to someone on the list.
The Saturday sheet is built the same way, its motive already settled above. The delivery platform stays out on purpose: nobody books a table from a delivery app, and announcing Thursday there means competing against your own dispatch.
Illustrative example generated with a fictional restaurant. Not a promise of results, not a client case.
This is 3 of 5 parts. Behind the wall sit the two pieces that get operated on any given Thursday. The full no-show mechanic: the policy written in two lines to send as is over WhatsApp, the release rule —confirmation Wednesday at 7:00 p.m., a second round Thursday at 1:00 and, with no answer by 4:00 p.m., the table goes to the list— and the four-column sheet, Table, Hour, Contact and Outcome, with who makes the call and within how many seconds to resell the released table. And the shift-opening checklist from 5:00 p.m.: the ten committed tables split between the two servers, pots counted against confirmed tables, the waiting list printed on the host stand, the 7:20 holding cut-off and the three figures read on Monday, starting with seated covers.
See the full example — free accountWorks with these AIs
How to use it
Which of your data it uses
This assistant works with 8 of the 23 fields in «My Restaurant», among them Strong and slow dayparts, Active sales channels 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.
v1.3.02026-08-25Card completed (doc 11): who it is for, a sample result with the fictional restaurant Sazon de Origen, and recommended AIs. The assistant body did not change.v1.2.02026-08-20Rooting: the TASK stops naming concepts and works with the house's own data. Peak slots decide which slot enters a campaign, team size caps how many tables open, reasons for visiting supply the motive, average check says whether the committed table is worth it, and sales channels decide where each piece goes out; the format writes the announcement for the main customer, the check compares price against the price range, and the next step counts covers against the city's calendar. Eight fields inside the working sections, five in the TASK, with no words added: two traps the steps now absorb were dropped.v1.1.12026-08-14Spelling: Spanish accents restored in the prose. No field marker, link or meaning changed.v1.1.02026-08-13Rewritten under the margin mandate: 6 numbered steps for the reservation book, 5 named output blocks with columns, 8 business fields, named traps, a verification checklist and confirmation windows with units.v1.0.02026-08-11The assistant is born.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
Designs a complete campaign end to end: money objective, audience, message, offer, channels and a 4-week calen…
N3 · Analysis with a verdictBuilds delivery campaigns on two lanes: selling more inside the platforms without melting margin into commissi…
N3 · Analysis with a verdictDesigns your loyalty program as a recognition ladder: steps with clear signals, house gestures that cost littl…
Included in the always-growing library
Access to every published assistant, adapted to your AI and personalised with your restaurant's data, updates included.
