Restaurant and café owners spend most of their day in the kitchen and at the till, while other work keeps waiting. Menu descriptions for the online store are still unwritten, customer chats about reservations pile up, and the sales reports in the POS rarely get a careful read. Promotions are often decided on instinct, simply because there is no time to look at the numbers.
AI for restaurants and cafés can help with each of these. This article covers realistic uses, the data they need, what to be careful about, and how to start from the systems you already use every day.
Writing menu descriptions
Good menu descriptions help customers choose, especially when they order through an online store or app and cannot ask a waiter. AI can write short descriptions from the details you give it, such as the main ingredients, cooking method, spice level, and portion size. It can produce them in Indonesian and English for cafés that see a lot of tourists.
The thing to watch is claims. The AI does not know your recipes unless you tell it. Do not let a description say "MSG-free", "gluten-free", or "halal" unless you have confirmed it. Allergen information such as peanuts, shrimp, and dairy must be checked by someone who knows the recipe, because a mistake here can harm a customer.
Replying to customer questions
The same questions arrive every day on WhatsApp and Instagram. What time do you open, can you take a booking for twenty people, is there a private room, do you deliver to a certain area, what do you offer for office events. AI can draft replies from a single official information document covering opening hours, delivery areas, capacity, and booking rules.
For reservations and catering orders, staff still confirm, because table availability and kitchen capacity change daily. Complaints are best answered directly by the owner or manager. AI can help polish the wording, but the apology and the offer to put things right are decided by a person.
Reading sales and stock data from the POS
Your cashier app holds very useful data: which items sell at which hours, which days are quietest, the average transaction value, and which dishes hardly anyone orders. The trouble is that nobody reads it, because the reports are long.
AI can condense POS reports into points that are easy to act on. For example, breakfast items sell well on weekdays but drop at weekends, or one drink has barely been ordered in the last two months. If ingredient stock is recorded in the system too, AI can flag items running low before the rush.
How accurate this is depends on how cleanly orders are entered at the till. If the same dish is recorded under several names, or ride-hailing orders are written in a separate notebook, the summary will mislead you. Tidying the menu list and entry habits is the first job.
Planning promotions
With sales data in hand, AI can help shape more targeted promotion ideas. A value set during quiet hours, a bundle of items that are often ordered together, or a coupon for customers who have not visited in a while. It can also draft promo copy for social media and messages to members.
The final decision is yours. Work out the margin on every promotion using your real cost of goods, since the AI does not know what you paid for ingredients this week unless that data is available. Review each promotion once it has run, and compare sales before and after.
Steps you can take this week:
- Clean up the menu list in your POS, including names, categories, prices, and variants.
- Write one official information document: opening hours, delivery areas, capacity, booking rules, and verified allergen information.
- Use AI to rewrite descriptions for your ten best-selling items, then check each one.
- Ask for a weekly sales summary and note one decision you made because of it.
- Test one small promotion in a quiet time slot for two weeks before rolling it out further.
Customer data and team habits
Restaurants keep customers' phone numbers for reservations, delivery, and membership programmes. That is personal data protected under Indonesia's Personal Data Protection Law (UU PDP, Law No. 27 of 2022). When you use a general AI assistant, there is no need to paste in your full customer list. For promo messages, ask for a template with placeholders like [customer name], then send it from the system that actually holds your member data.
Agree on how the team uses AI as well. Keep prompts that work well in one shared document. Decide who checks text before it goes public, for example the outlet manager for menu descriptions and the owner for promo material. AI is also most useful when it works with data recorded properly in a system, rather than numbers copied into a chat by hand.
Talk it through with Averoa
Etalasse combines an online store and a POS cashier with table management and a kitchen display, plus an AI assistant that helps manage products, stock, and orders by chat. Its loyalty tools, such as coupons, points, memberships, and flash sales, let you run the promotions you have planned. If you would like to see whether this way of working suits your restaurant or café, the Averoa team is happy to talk it through. See the full feature list on the Etalasse page.
