Luca, an AI waiter to help your guests choose.

A guest does not always start with the name of a dish. They may have a budget, want a particular category or simply need somewhere to begin. Luca connects these questions to products that exist in your menu. Recommendation cards show recorded product names and prices, and guests can continue exploring the details in the menu.

For example, a question such as “What is available under $20?” uses the budget the guest has supplied and the menu’s recorded prices. This amount is an example of a guest request, not an Orvunai subscription price or a sample restaurant offer. If the menu has no suitable option, the experience should make that boundary clear instead of inventing a cheaper dish.

Luca’s product choices are restricted to published, visible and available items. Prices come from the menu rather than a guess. The quality of your content therefore matters: clear categories, useful descriptions and current prices support both ordinary browsing and discovery through the assistant. Keep the menu organized around what your kitchen actually serves.

A product you mark as featured reflects your restaurant’s choice. A guest favorite is a recorded interaction. Neither should be confused with a sales count, and an assistant recommendation is not a completed order. Guests still communicate their choice through your existing service process; the conversation does not send a ticket to the kitchen.

When products have restaurant-verified dietary labels that do not conflict with other recorded information, those labels can support discovery. A dish you have verified as vegetarian, for example, can be considered for that preference. An empty field does not prove that a product has, or lacks, a particular property.

Allergies, medical needs, unknown ingredients and missing nutrition information require a different response. Luca does not infer those facts from a dish name and directs guests to your team. Your staff must confirm recipes, preparation methods and possible cross-contact. A conversation cannot establish conditions that have not been verified by the restaurant.

Assistant settings let you enable or disable the experience and edit its displayed name, welcome message and tone setting. Luca is the name used in Orvunai’s own introduction; you can choose a name that suits your restaurant. A useful welcome stays short and gives guests a clear example of how to begin.

“Welcome. Tell me a category or budget to get started” sets a practical expectation. Avoid adding promises that your kitchen cannot verify, such as offering to find a safe meal for every allergy. Explain that the assistant helps guests explore available menu products, with your team ready to answer questions that need direct confirmation.

After publishing, open the assistant as a guest. Try different budgets, a product category and something that is not on your menu. Check the names and prices on the returned cards. When product availability changes, verify the menu and assistant experience so guests are exploring options that match the current service.

The homepage conversation is an example demonstration. The live assistant is not an unlimited chatbot, an ordering service or an automatic food-pairing engine. It does not infer which dish should accompany another or how many people a serving will feed without an authoritative basis. Assess the product through its actual menu discovery behavior.