ChatGPT prompts for restaurants: 16 copy-paste prompts built from what real restaurants do (2026)

What are good ChatGPT prompts for a restaurant menu and pricing?

Start with the dish, then the numbers. Nina Zhang asks ChatGPT whether American diners will like a traditional dish before she puts it on, and had it come up with the "golden lava bun" for her yolk-custard bun. Ted Taylor gives ChatGPT Work a categorized menu with in-store and delivery prices and the increase he wants, and gets back a spreadsheet of recommended prices. For events, Christie's works out quantities and costs so it makes less food it throws away.

I run {RESTAURANT}, a {CUISINE} restaurant in {CITY}. Here is a dish we want to sell more of: {DISH DESCRIPTION, ORIGINAL NAME, KEY INGREDIENTS}. 1) Tell me honestly what might make diners here hesitate to order it and what might appeal to them. 2) Suggest 10 English menu names (5 plain and descriptive, 5 playful), each with a menu description under 20 words. Don't invent ingredients or health claims, and don't call it authentic or traditional unless I say so.

Attached is my menu for {RESTAURANT} in {CITY}: item, category, in-store price, delivery price, plate cost and units sold over {PERIOD}. Recommend new in-store and delivery prices so the overall increase is about {PERCENT}% and food cost stays near {FOOD COST TARGET}%. Rules: don't change {ITEMS TO HOLD}, keep delivery prices {DELIVERY MARKUP} above in-store, and avoid any single increase over {MAX INCREASE}. Return a spreadsheet with old price, new price, change, new food-cost % and a one-line reason, and flag the items where a big increase could cost me regulars.

Using my portion and cost sheet, plan food for an event at {RESTAURANT NAME}: {GUEST COUNT} guests, {SERVICE STYLE, e.g. buffet or plated}, menu: {MENU ITEMS}, event length {HOURS}. Give a table with each item, portion per guest, total quantity to prep, raw ingredients to order, and cost. Add the total food cost and cost per guest. Show the math, and note where you assumed something not in my sheet.

What ChatGPT prompts help with restaurant inventory and the kitchen?

Inventory is where a photo or a spreadsheet beats typing. Cole Morgan snaps the café's paper inventory sheet and gets a checklist of what to buy, grouped by store. Los Dos Potrillos compares what sold with what was left to see whether it orders too much alcohol, rice or beans. When equipment breaks, Ramy Singh photographs it and asks AI what's wrong before calling a repair tech, and Brewpoint Coffee's owner uploaded all its equipment manuals into ChatGPT for troubleshooting.

You are the remote inventory manager for {CAFE NAME}. When I upload a photo of today's inventory sheet, compare it with the item list in the project file and output only a markdown checklist of items to buy, with checkboxes, grouped by vendor ({VENDORS, e.g. Sams Club, Hannaford}). Before finalizing, ask me about any item on the list that isn't on the sheet. Don't add items that aren't on the list.

Here are two weeks of data for {RESTAURANT}: item sales from our POS {SALES CSV}, what we ordered from suppliers {ORDERS}, our end-of-week counts {COUNTS} and the amount of each ingredient per dish {RECIPE AMOUNTS}. For {INGREDIENTS e.g. rice, beans, tequila, beer}, calculate expected use, actual use and what we have left. Flag anything where we ordered more than {PERCENT}% above what sold, and suggest a new weekly par for each. Show the math and mark any assumptions.

These photos show our {EQUIPMENT: FRYER / OVEN / ICE MACHINE}, model {MODEL FROM DATA PLATE}. Symptom: {WHAT IT DOES OR DOESN'T DO}. Error code shown: {CODE OR 'none'}. List the most likely causes in order, and for each say (1) what a staff member can safely check with the unit off and cool, (2) what needs a licensed technician, and (3) which page or section of the manufacturer's manual to look for. If you can't read something in the photo or aren't sure of the model, say so instead of guessing. Remind me of any gas or electrical safety steps first.

Using only the attached equipment manuals, help me diagnose: {SYMPTOM} on {MACHINE}. List likely causes in order, the checks to do, and when to call a technician.

What ChatGPT prompts help a restaurant with sales reports, P&Ls, payroll and cash?

Upload the export you already have. Old Scratch Pizza runs a fixed prompt on a weekly Toast Order Details CSV (over 8,000 lines) to get the share of checks done in 20 minutes or less at each location. Vice Beer's co-founder asks ChatGPT to act as CFO on monthly P&Ls. Cowbell checks weekly payroll summaries against set benchmarks and a checklist, and Ted Taylor has each store's cash matched to bank deposits every month. Have your bookkeeper or accountant sign off on anything that gets booked.

Analyze restaurant order fulfillment times in the attached Toast Order Details CSV for {DATE RANGE}. Group by Location, Check # and Date. For each check, take the earliest Fired Date to the latest Fulfilled Date. For each location show: total checks, number and % fulfilled in {TARGET MINUTES, e.g. 20} minutes or less, and number and % over. Format percentages to two decimals. Then make a side-by-side bar graph comparing locations by day of week.

Act as the CFO of {BUSINESS NAME}, a {TYPE, e.g. small brewery with a taproom and wholesale accounts} in {CITY}. Review the attached P&L statements for {MONTHS} and any sales or production data. List the three biggest trends, any line where we're spending more than {BENCHMARK, e.g. payroll over 30% of sales}, and decisions I should make this month. For any new product idea I mention ({IDEA}), list the costs and what it ties up (cash, kegs, tank or shelf space, staff time) and for how long. Quote the figure you're using for every point.

Check this weekly payroll summary for {RESTAURANT NAME} against these benchmarks: {BENCHMARKS, e.g. labor under 32% of sales, no one over 40 hours without approval, tip pool split per our tip model}. Go through this checklist: {CHECKLIST ITEMS}. List every line that breaks a rule, with the employee role, the number and the rule. Don't change any figures; if data is missing, say so.

Attached are {MONTH} cash records for {STORE}: daily cash sales and deposit log from our POS {POS CSV} and the bank statement {BANK CSV}. Match each day's cash to a bank deposit (deposits may land 1-3 days later or combine days). Return a table: business date, cash expected, deposit date, amount deposited, difference. Then list every unmatched deposit, every day with no deposit and every difference over {DOLLAR THRESHOLD}. Don't guess at causes; just flag them.

Can ChatGPT build a restaurant staff schedule?

Two restaurants here have it draft one. Tony Mistry of BnX Burgers uploads past schedules and the week's availability changes to paid ChatGPT and gets updated shifts. Ted Taylor's ChatGPT Work drafts one store's full schedule from the roster and staffing rules; the manager checks it and enters it into the scheduling system. Both keep a person on the final version.

Build next week's staff schedule for {RESTAURANT NAME}, open {HOURS}. Use the uploaded past schedules as the pattern. Staff, roles and limits: {STAFF LIST WITH ROLES, MAX HOURS, DAYS OFF}. Changes this week: {CHANGES}. Minimum coverage: {COVERAGE, e.g. 2 cooks and 2 front-of-house Friday 5-9 PM}. Output a table by day and shift, total hours per person, and flag anyone over {OVERTIME LIMIT} hours or any shift below minimum coverage.

Here are 12 weeks of data for {STORE}: daily sales (forecast and actual) {SALES}, hours by role (scheduled, budgeted, actual) {HOURS}, and our labor target of {LABOR %} of sales. 1) Flag overtime, shifts that ran longer than planned and under-covered hours, and say which store or day needs attention. 2) Using this roster {ROLES, TRAINING, AVAILABILITY, MAX HOURS} and these staffing rules {RULES}, draft next week's schedule for forecast sales of {FORECAST}. Show projected labor % per day and list every rule you couldn't meet.

What ChatGPT prompts work for restaurant flyers, review replies and customer emails?

Feed it the real thing: a photo of your actual food (why that matters), examples of your own voice, and your real prices and policies. Jasmin Walls makes flyers from real photos of her donuts, Old Scratch Pizza uploaded examples of its brand voice before drafting review replies, and Sara Rojas pastes in what a client sent and gets back a draft reply in the shop's voice, which she tweaks and sends herself.

Here is a real photo of {ITEM} from {BUSINESS NAME}. Make a {SQUARE / STORY} social media flyer for {OFFER} on {DATES} at {PRICE}. Keep the food exactly as it looks in my photo: don't change its shape, color, size or toppings, and don't add food that isn't there. Text on the flyer, spelled exactly: "{BUSINESS NAME}", "{OFFER}", "{DATES}", "{ADDRESS}". Use no other text. Then write a 2-sentence caption in a warm, family-run voice.

You write for {RESTAURANT NAME}, a {TYPE OF RESTAURANT} in {CITY}. Match the voice in the uploaded examples. When I paste a review, draft a reply under {WORD LIMIT} words that thanks the guest by first name, mentions one specific thing they said, and, if they had a problem, apologizes once and invites them to contact {CONTACT}. Never offer refunds or freebies unless I say so.

You help me answer customer inquiries for {BUSINESS_NAME}, which sells {WHAT_YOU_SELL} in {CITY}. My tone: {TONE_RULES, e.g. warm, friendly, short, never robotic}. Facts you can use: {PACKAGES_PRICES_AND_POLICIES}. Here is the client's email: {CLIENT_EMAIL}. Draft a reply that answers each question they asked, asks for any details I need (date, guest count, budget), and ends with a clear next step. Do not invent prices or availability. Mark anything you are unsure of with [CHECK].

What should restaurant owners watch out for with ChatGPT?

• Mistakes in numbers and text. Drazick Halip calls it "kind of like a stage or an intern." Check totals against your POS and read every word on a flyer before it goes out.
• It guides, you decide. "We like to let AI guide our thoughts, not think for us," says Daniel Ramirez of Los Dos Potrillos; Taylor reviews results before acting on them.
• Private data. As a suggestion, mask bank account numbers, leave out employee details the job doesn't need, and check your ChatGPT plan's data settings before uploading payroll or bank files.
• Legal and safety calls. For trademarks, health rules or equipment repairs, use the AI's answer as a starting list and confirm with the professional, the agency or the manufacturer.