AI assistant · N5 Complete system
AI Costing Spreadsheet Analyzer for Restaurants
Hand your costing spreadsheet to the AI and receive the tool: a report on the dishes bleeding margin, errors in the sheet itself, and an analysis file ready to rerun every month.
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
- Almost every restaurant has ONE costing spreadsheet — and almost none has audited it.
- This assistant, pasted into an agent that can read and create files (Claude with Cowork or Claude Code), takes your recipe-and-cost sheet as it is, audits it first (broken formulas, impossible portions, unpriced ingredients, wrongly computed margins) and then analyzes it with the house criteria: real food cost per dish against the ceiling you declare —and if you have none, the assistant writes its assumption in plain sight instead of inventing an industry figure for you—, contribution margin in money and not just percentage, and the double ranking of what to push and what to reformulate.
- Finally it builds you a reusable analysis file so costing stops being an event and becomes a monthly routine.
What you end up with
Three pieces: the audit of your sheet with each finding and its cell pinpointed; the costing report with real food cost per dish, contribution margin in your currency, the double ranking (push / reformulate) and this week's 5 decisions; and a ready analysis file (analisis_costeo.html with instructions inside) to rerun the reading every month.
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: «Here is my costing spreadsheet (42 recipes, 3 sheets). Tell me where it is wrong and which dishes are eating my margin.»
The sheet first, the menu second: a flawless analysis built on a broken formula is a lie with formatting.
| Finding | Cell or row | Severity | Proposed fix |
|---|---|---|---|
| Beef loin waste is set to 0% and the real cost lands too low | Recipes!D14 | Breaks the analysis | Declare the waste measured across three butcherings |
| 6 recipes sum with ranges that miss the new rows | Recipes!F22:F27 | Breaks the analysis | Extend the range and lock the totals row |
| The coastal cheese price has not moved since January | Supplies!C31 | Distorts | Stale price: over 90 days with no new date |
| 3 recipes use portions in grams and purchases in kilos, unconverted | Recipes!E8, E19, E33 | Distorts | Unify the unit in the purchase column |
| Sheet «old2» duplicates 11 recipes and skews the averages | whole sheet | Cosmetic | Archive the sheet outside the live workbook |
SUPUESTO: no industry food cost ceiling is hard-coded here. The ceiling used is the one the house declares on its card; until it does, the traffic light compares against the sheet's own historical food cost, and says so. If the real ceiling were 3 points lower, two of the green dishes would turn amber.
Reading window: the last 4 weeks of sales, with the real POS mix.
| Dish | Real cost per portion | Menu price | Food cost | Unit margin | Month's margin | Verdict |
|---|---|---|---|---|---|---|
| Seafood cazuela | $19,800 | $52,000 | 38.1% | $32,200 | 1.9 M | 🔴 reformulate the portion |
| Lomo al trapo | $15,100 | $48,000 | 31.5% | $32,900 | 4.1 M | 🟢 push |
| House rice | $8,900 | $34,000 | 26.2% | $25,100 | 5.6 M | 🟢 push |
| Sea bass in sauce | $16,700 | $46,000 | 36.3% | $29,300 | 1.2 M | 🟡 renegotiate the supply |
The ranking is double on purpose. By percentage the rice would win and the cazuela would lose; by month's margin the rice still wins — but the loin, with a worse percentage than the rice, contributes more money than the cazuela and the sea bass together. With the waste corrected, the month's measured food cost lands 2.5 points above the declared one: that food cost variance is what the sheet was hiding.
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 are **this week's five decisions** —dish, owner with job title, deadline and the figure that confirms each one— and **the analisis_costeo.html file** with its zone for pasting next month's reading.
See the full example — free accountWorks with these AIs
How to use it
Which of your data it uses
This assistant works with 9 of the 23 fields in «My Restaurant», among them Restaurant name, Business type and Number of locations. 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.
v1.2.02026-08-25Body rewritten to pay down duelo debt: it scored 59.4 out of 90.0. The `metodo` signal read ZERO because the three chained deliverables lived inside one running paragraph («Deliverable 1 — … Deliverable 2 — …»), and `verificacion` read 1 because BEFORE YOU DELIVER was a row of questions without a single checking verb. It is now six numbered steps at the start of the line that inject «fields» into the TASK (required by arraigo, which measured 2 fields in the work sections and 1 in the TASK), plus five imperative checks. The 32% food cost ceiling that was hard-coded as industry truth is removed: the card now orders the assistant to declare its assumption with the SUPUESTO: line and state what would change in the verdict if the real value were different. `limites_es` and `limites_en` are added because the title says «with AI» and alcance.py requires the card to state that it does NOT read the POS nor hand back a corrected Excel. Rises to 95.0.v1.1.02026-08-14Card expanded as a lead magnet (doc 11): who it is for, a sample result with the fictional restaurant Sazón de Origen and recommended AIs with an honest note on agent delivery. The assistant body did not change.v1.0.02026-08-14Born in the TOOLS line (Diego's 2026-08-14 commission): assistants that read and build files inside an agent.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
Builds your restaurant's cost board as an interactive file: food cost, prime cost, waste and purchasing by per…
N3 · Analysis with a verdictMeasures your bar cost split by family, bottled drinks, beer, wine, cocktails and coffee, shows which family l…
N3 · Analysis with a verdictRebuilds each dish's real cost at today's purchase prices, computes cash margin and price floor per item, and …
Included in the always-growing library
Access to every published assistant, adapted to your AI and personalised with your restaurant's data, updates included.
