Who is this assistant for?

You have three new dishes ready in the kitchen and the menu goes to print on Friday. Your server explains the same dish at every table because the name does not say what it is. Nobody finds your dish on the delivery app because you gave it a pretty name. You want to rename a dish with years behind it and do not know if you lose more than you gain.

Naming a dish is a different job from describing it: the name is read in a second, fits one menu line, gets said out loud at a table and gets typed into a delivery app's search box. This assistant covers both directions —christening a new dish and rescuing one that does not work— with discard criteria you can verify: pronunciation, comprehension, collision with another name on the same menu, and a promise the dish does not keep. And it makes, in cold blood, the call almost nobody makes: when NOT to change a name that already has years of repeat orders. If what you need is the description that sits under the name, that is the neighbouring card, the Menu Description Rewriter.

AI assistant · N4 Executable plan

◆ Included in the annual plan🍽️ Menu Engineeringv1.0.1 · 2026-08-26

Dish Name Builder for Restaurants

Names your new dishes and rescues the ones nobody orders: twelve candidates, four verifiable discard criteria and the final name written for the menu, the kitchen docket, the point of sale, the delivery listing and the server's mouth.

Depth

10/23

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 ▶

🗺️

N4Executable plan

A plan with owners, deadlines and the metric you review. Not advice: an agenda someone can execute on Monday.

◆ This assistant 23 of the 164 published assistants are at this level

Is this for you?

What it does and what you get

Naming a dish is a different job from describing it: the name is read in a second, fits one menu line, gets said out loud at a table and gets typed into a delivery app's search box.
This assistant covers both directions —christening a new dish and rescuing one that does not work— with discard criteria you can verify: pronunciation, comprehension, collision with another name on the same menu, and a promise the dish does not keep.
And it makes, in cold blood, the call almost nobody makes: when NOT to change a name that already has years of repeat orders.
If what you need is the description that sits under the name, that is the neighbouring card, the Menu Description Rewriter.

What you end up with

Five chained deliverables: each dish's card with its entry decision, twelve candidates with their quarry, the sifting with the written reason for every discard, the name ready for its four places and the changeover calendar with a 90-day cut-off.

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: "I have three new dishes with no name and an old one, the Plato del chef, that almost nobody orders. I want to name the three and decide what to do with the fourth before the menu goes to print."

Five dishes on the table and a different decision for each one. Before a single name was proposed, every dish got its card and its costing sheet: a name that does not come out of there is a bright idea in nice handwriting.

Dish	What it carries	Name today	Menu price	Cost per plate	Food cost	Entry
Panela-glazed pork rib	rib, cane-sugar panela, arracacha purée	no name	46,000 COP	14,720 COP	32%	Christening
Young-corn parcel	choclo, fresh quesito, husk	no name	28,000 COP	9,520 COP	34%	Christening
Oven-baked flank	sobrebarriga, criolla potato, hogao	no name	52,000 COP	17,160 COP	33%	Christening
Rice with chicken, chorizo and plantain	rice, chicken, chorizo, plantain	"Plato del chef"	38,000 COP	13,300 COP	35%	Rescue
Traditional platter	beans, chicharrón, arepa, egg	"Bandeja de la casa"	54,000 COP	18,360 COP	34%	Hands off

"Plato del chef" is the one the house wants to push, and the finding there is not about the name: at 35% food cost it sits above the 32% the house sets as the ceiling for a promoted piece, with the menu averaging 34%. At 41,600 COP the same dish lands at 32%; if the price does not move, cost per plate has to come down to 12,160 COP. It gets named anyway, but it is not promoted until one of the two gives: the problem there is not the name, it is the menu price.

"Bandeja de la casa" is not touched, and that was the hardest call of the session. Six years on the menu, 22% of units sold over the last 12 weeks, and guests who order it without opening the menu. A new christening would have to rebuild that recognition from zero, and recognition never shows up on a costing sheet even though it gets paid for. Illustrative figures, built from the data you gave.

Four quarries produced twelve candidates. Here are the four chosen and the four discards that teach the most; the other four sit in the full sieve.

Candidate	Quarry	Characters	Understood without asking	Collision	Verdict
Costilla en panela	Anchor ingredient	18	Yes	None	CHOSEN
Envuelto de choclo	Anchor ingredient	18	Yes	None	CHOSEN
Sobrebarriga al horno	Technique	21	Yes	None	CHOSEN
Arroz de la fonda	The house	17	Yes	None	CHOSEN for the rescue
Costilla braseada a baja temperatura	Technique	36	Yes	None	Out: does not fit the 24-character field on the point of sale
Panela y candela	The house	16	No	None	Out: sounds good and says nothing about the food
Costilla en su jugo	Anchor ingredient	19	Yes	None	Out: promises a sauce the dish does not have
Sobrebarriga de la casa	The house	23	Yes	Clashes with "Bandeja de la casa"	Out: "de la casa" already has an owner on this menu

Two discards carry the whole method. "Costilla braseada a baja temperatura" was the most precise of the twelve and fell over 12 characters too many: the point of sale cuts it, the server types whatever fits, and the sales mix stays dirty for the rest of the year. And "Costilla en su jugo" promises sauce where there is a dry glaze; that promise comes back as a returned plate, and with that rib's food cost at 32%, two send-backs wipe out the table's contribution margin.

The collision rule is not about taste: "Sobrebarriga de la casa" and "Bandeja de la casa" abbreviate the same way on the docket, and an ambiguous docket at 1:10 p.m. on a Thursday costs one plate remade.

The chosen name, written for each place where it lives. Four places and five rows, because this house's delivery channel is two: its own and the delivery platform. The character caps are assumed until you check them against your point of sale.

Place	How it is written	Character cap	Who types or says it	Signal measured
Dining-room menu	Costilla en panela	18, one line	Admin	Units sold per week
Docket and point of sale	COSTILLA EN PANELA	24 (assumed)	Server	Dockets with no handwritten note
Own delivery, over WhatsApp	Costilla en panela	free	Delivery assistant	Orders that name it without seeing a photo
Delivery platform	Costilla en panela · cerdo glaseado con arracacha	60 (assumed)	Admin	Searches that end in an order
Server's mouth	"la costilla en panela, that bakes for 12 hours"	one sentence	Floor manager	Tables that order it without asking what it is

On the platform people type "costilla", not "panela": that is why the searched word goes first and the explanation comes after the middle dot, where the title still fits. In the dining room the short name is enough, because the description sits below it. In the server's mouth the name travels with one kitchen fact and no more: if it needs two, the name was not finished.

The same 18 characters hold all five rows without breaking. That is the whole test, and it is run before the menu goes to print, not after.

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 parts that get executed. The full sieve, with the twelve candidates —Candidate, Quarry, Characters, Understood without asking, Collision, Promise kept and Verdict—, the four that do not appear above and the written reason for every fall; and the script read to five people who do not know the dish, with its rule: a name passes when 4 of 5 repeat it correctly and 4 of 5 get the protein right. And the changeover calendar, from week −1 to week +12, with move, place, owner and signal on every row, the old name coming off the platform listing in week +4, and the cut-off read on Monday at 9 with the sales mix: if the new name does not reach 100 units a month, the old one comes back.

See the full example — free account

Works with these AIs

ChatGPT — the best for this assistant The best combination for this assistant: it produces candidates in volume, counts their characters and writes the server's line without switching register in between.

How to use it

Have the standard recipe card and the costing sheet of every dish you plan to name, plus the current name and its units sold if the dish already exists.

Copy the assistant into your preferred AI along with your business details and answer its questions one at a time, about the menu and about your point of sale.

Measure the real character cap of your point of sale and of your delivery listing before you approve the final name.

Put the chosen name in all four places on the same day, write down the date and read the cut-off at 12 weeks with the sales mix in hand.

Which of your data it uses

of 23

This assistant works with 10 of the 23 fields in «My Restaurant», among them Restaurant name, Cuisine or concept 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.

Restaurant nameCuisine or conceptCity and countryYour main guestBest-selling dishesDish descriptions (paste your menu)Active sales channelsAverage check (with currency)Menu price rangeWhat makes you different