Who is this assistant for?

The sheet's photo is from the photo shoot and today's plate carries a different garnish. You suspect the weight has drifted, but everyone tells you something different. You have no idea what your floor team is saying about the dishes or how they answer allergen questions. You changed menu prices and cannot say how many sheets were left carrying the old one.

You have spec sheets from when the menu was built and the plate that goes out today looks nothing like the photo. This assistant audits during service rather than at a desk: it weighs what leaves, compares the photo, asks the floor what they know and opens a delivery box. It returns a verdict per sheet —faithful, drifted or expired— and orders corrections by risk, starting with allergens rather than with whatever is quickest to fix.

AI assistant · N5 Complete system

◆ Included in the annual plan👨‍🍳 Standard Recipes, Spec Sheets and Productionv1.0.0 · 2026-08-26

📷 Works best with an AI that sees photos (invoices, dishes, screens)📚 Works best with an AI that reads long documents in one pass (your full menu, a manual) · optional

Dish Spec Sheet Auditor for Restaurants

Holds every spec sheet against real service with four tests —the weighed plate, the photo, the server and the delivery box— and returns a verdict per sheet with corrections ordered by risk.

Depth

9/23

Your data

Compatible AIs

Steps to use it

Depth of the resulttap a level to see what you get

◀ Put out today's fireRun the year ▶

🏛️

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 level

Is this for you?

What it does and what you get

You have spec sheets from when the menu was built and the plate that goes out today looks nothing like the photo.
This assistant audits during service rather than at a desk: it weighs what leaves, compares the photo, asks the floor what they know and opens a delivery box.
It returns a verdict per sheet —faithful, drifted or expired— and orders corrections by risk, starting with allergens rather than with whatever is quickest to fix.

What you end up with

The sheet inventory against the current menu with a verdict and a risk per dish; the weight deviation table component by component with what today's photo has and the sheet's does not; the floor's verbatim answers with the allergens nobody named; and the correction plan ordered by risk with owner, date and a faithful-sheets indicator.

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

Sazón de OrigenFictional contemporary Colombian restaurant in Medellín, 2 locations, 6 years in business. Every sample result in the library comes from this same case. Fictional restaurant

The owner asked this: "We built the spec sheets when we opened the second site and we have changed prices twice since. I want to know which ones still work before printing the new menu."

Dish	Sheet	Date	Sheet price vs menu	Plate test	Server test	Box test	Verdict	Risk
House bandeja	yes	03-2025	39,000 vs 42,000	−8.2 % on beef	2 of 2 named the dish, 0 of 2 the allergens	hogao travels on top, not apart	EXPIRED	PRICE
Three-meat sancocho	yes	03-2025	same	+11.0 % on cassava	1 of 2 could describe it	no delivery note	DRIFTED	WEIGHT
Trout with garlic	yes	03-2025	51,000 vs 54,000	within tolerance	0 of 2 mentioned it carries butter	box correct	EXPIRED	ALLERGEN
House rice	no	—	—	—	—	—	NO SHEET	WEIGHT

Of 12 sheets audited: 3 FAITHFUL, 5 DRIFTED, 3 EXPIRED and 1 menu dish with no sheet at all. The finding that orders the whole plan is not the biggest one, it is the gravest: the trout carries butter and neither of the 2 floor staff mentioned it. That does not get fixed with a meeting: it gets fixed today on the sheet and tomorrow at pre-shift.

House bandeja · 5 plates at random, Thursday 13:10-13:55, with no warning to the pass.

Component	Declared weight	Weighed (mean of 5)	Deviation	In percent
Cooked beans	210 g	232 g	+22 g	+10.5 %
White rice	180 g	174 g	−6 g	−3.3 %
Shredded beef	95 g	87 g	−8 g	−8.4 %
Chicharrón	55 g	51 g	−4 g	−7.3 %
Hogao	45 g	62 g	+17 g	+37.8 %

What this table says, read properly. Beans and hogao go up, protein comes down. That is not carelessness: it is what happens when weight rests on a hand rather than on a document, because the cheap thing gets served generously and the expensive thing gets served prudently. The plate still weighs about the same —the guest notices nothing on a scale— but the contribution margin moved without anyone deciding it, and the 48,000 COP average check is being propped up with extra beans.

What the sheet's photo has and today's plate does not: the lime wedge on the rim, dropped in March for cost with nobody updating the image. What today's plate has and the photo does not: the hogao served on top of the beef instead of beside it, which is why it arrives mixed in delivery.

SUPUESTO: the tolerance these deviations are judged against is the ± 10 g written on the house's own sheet, not an industry norm. Had that tolerance been set at ± 5 g, four of the five components would fall out of range and the verdict would move from DRIFTED to something harsher. Before arguing about the verdict, argue about the tolerance: it is the yardstick, and the house set it.

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 is **the part no owner can audit alone**: the verbatim answers of the two floor staff, the list of allergens nobody named crossed against supplier and sheet, the matrix of substitutions the floor believes are allowed, the delivery box opened at 25 min against the table plate, the plan by risk with allergens closed within 48 hours, the re-audit checklist and the indicator with its fixed day.

See the full example — free account

Works with these AIs

Claude — the best for this assistant The best one here because the grave finding gets buried among the many: it tells the allergen nobody named from a weight that drifted three points, and it orders by risk even when you ask for the report ordered by dish. Give it both photos together and ask it to say what it sees in each before it judges.

How to use it

Fill in My Restaurant and gather whatever spec sheets exist, with the current printed menu beside you.

Attach a photo of the dish as it went out today and the photo on the sheet, then paste the assistant into your preferred AI.

Run the plate test at peak time without telling the team, five plates at random and a scale.

Interview two floor staff separately and write down what they say exactly as they say it.

Close the findings in order of risk and re-audit the whole set every 90 days.

Which of your data it uses

of 23

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.

Restaurant nameBusiness typeNumber of locationsBest-selling dishesDish descriptions (paste your menu)Menu price rangeActive sales channelsYour main guestAverage check (with currency)