AI assistant · N5 Complete system
Operating Manual Builder by Role for Restaurants
Pick the role and your team size above, and the assistant writes its whole shift hung off the clock: five stretches with numbered sequences, a standard with number and unit on every step, the check, the escalation, the ten most frequent failures and the handover script. Twenty-nine loose manuals in one with a dropdown.
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 operating manual nobody uses is the one written to be read cover to cover.
- This one is written to be consulted at 12:40 with both hands busy: hung off the shift clock rather than the org chart, in numbered steps, with the standard in degrees, grams, minutes or cash beside every step, and with the check that says whether it came out right — not whether it was done.
- It is the document that lets the shift repeat itself when you are not there, that gets a new cook to standard in days instead of months, and that turns the handover into something more than two sentences at the door.
- It does not define the role — that is the role manual, a different document: this is how the work is done, step by step.
What you end up with
The operating manual for the chosen role: the shift clock as a table with its five stretches and durations; one block per stretch with the numbered sequence — twelve steps at most — and, beside every step, its standard with number and unit, what the check compares against, and the escalation with recipient, channel and minutes; the ten-failure table with cause, fix and cost; the five moments of truth with the exact phrase; the handover script in eight lines; and the fifteen-minute floor test to verify somebody can work it without it in front of them.
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: «I have high server turnover: a new one every two months, trained by whoever is standing next to them, so each one runs the shift differently. I need the server's manual for the Tuesday-to-Friday lunch, which is when we fill up.»
Five stretches, 42 minutes of your own work outside service; the rest is service.
| Stretch | Time | Duration | Aim of the stretch |
|---|---|---|---|
| Arrival and uniform | 11:00 | 10 min | Be ready before set-up, not during it |
| Set-up and pre-check | 11:10 | 20 min | Floor serviceable at 11:30, not 11:55 |
| The peak | 12:00–14:30 | 150 min | Hold ticket times and sales per table |
| Lull | 14:30–15:15 | 45 min | Restock, close open tabs, set up the evening |
| Close and handover | 15:15 | 12 min | Whoever comes in has nothing to ask |
If the peak starts at 12:30 at the other location, the whole clock shifts 30 minutes: you do not rewrite the manual, you shift the time column.
| # | Step | Standard | Check | Escalation |
|---|---|---|---|---|
| 1 | Greet and hand over the menu | Within 2 minutes of them sitting | Look at the ticket clock, not your impression | Past 4 min, radio the captain, immediately |
| 2 | Take the drink order before the food order | Drink on table within 5 minutes | Drink served before taking the main | Bar backed up: tell the captain before promising a time |
| 3 | Offer the starter of the day by name, not «anything to start?» | 1 specific suggestion per table | Count how many tables took it at close | If nobody takes it in 10 tables, say so at handover |
| 4 | Enter the ticket with modifications written, never spoken | 0 verbal tickets to the kitchen | The printed ticket says what the guest ordered | A kitchen error from a bad ticket is yours: flag it there and then |
| 5 | Check the plate before you carry it | Hot food leaves above 63 °C; no plate leaves without its side | Compare against the recipe-card photo at the pass | Off standard: back to the pass, not to the table |
| 6 | Return to the table 3 minutes after the first bite | 1 pass per table per course | Ask about the dish, not «everything okay?» | A complaint: call the captain before offering anything |
No step says «politely». Every one says what it is checked against.
| Failure | Likely cause | One-line fix | What it costs |
|---|---|---|---|
| Table seated ungreeted for 4+ min | Nobody owns that zone | Zones written at set-up, with names | The first impression, and the tip |
| Verbal ticket to the kitchen | Rushing at the peak | Zero exceptions: if it is not printed, it does not exist | A refired plate ≈ 1 average check |
| No starter or dessert offered | Fear of «selling» | One specific suggestion by name, not a question | Up to 12 % of the check |
| Lukewarm plate at the table | Food piling up at the pass | Leaves two plates at a time, not four | A return and waste |
Extract: the full table runs ten rows, ordered by same-day cost.
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 are the two pieces that make the manual something you can actually demand. The five moments of truth, with the exact phrase said at each one — the greeting, the suggestion, the arrival of the main, the complaint and the bill — written from the house's promise and tone of voice rather than a generic script; plus the handover script with the six things told in order and the two written in the book. And the full floor test: the timed exercise, the table with its two metrics — moments of truth recalled out of five and peak steps run in order out of eleven —, the 4-out-of-5 and 80 % cut-off below which nobody is left to run the shift alone, and the verdict box with date and assessor.
See the full example — free accountWorks with these AIs
How to use it
Which of your data it uses
This assistant works with 16 of the 23 fields in «My Restaurant», among them Restaurant name, Business type and Cuisine or concept. 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.
v1.1.02026-08-27The `tamano_del_negocio` axis comes in, Diego's exact brief: «it must take into account the size of the restaurant or the number of employees». It is the axis that changes a role manual the most: up to 10 people one role is covered by several people and one person covers several roles, so the manual says what is done and who covers it TODAY and not an org chart; above 30 there are middle managers and it must say who decides without calling the owner.v1.0.12026-08-27The response-ladder doctrine is stated in English with the house phrase —«continue with a marked assumption»— and not with a synonym. The migration probe caught it: 34 migrated bodies in Spanish and 27 in English, and the seven-card gap were the ones I wrote today. A doctrine said two ways stops being greppable.v1.0.02026-08-27Born by merging the twenty-nine «Operating Manual for the ___» cards in c27 — general manager, administrator, executive chef, head chef, sous chef, line cook, assistant, server, captain, host, cashier, bartender, barista, steward, storekeeper, delivery rider, accountant and the rest — which were the `cargo` axis written twenty-nine times. The role is picked with the selector and it changes the clock, the steps and the units.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.
Included in the always-growing library
Access to every published assistant, adapted to your AI and personalised with your restaurant's data, updates included.
