What Is AI Visibility? How ChatGPT and Gemini Recommend Restaurants

Picture this scenario: A family is planning a dinner out. Their eleven-year-old has a diagnosed peanut allergy. The father types into ChatGPT: „Which Chinese restaurants in Frankfurt offer peanut-free dishes?“ ChatGPT responds instantly with two concrete recommendations — complete with address, featured dish and the note that structured allergen data is available. Your restaurant is not among them. Not because your food is worse. But because your data is invisible to ChatGPT.

That is AI visibility — or rather, its absence.

→ AI Visibility Hub — Overview

The New Reality of Restaurant Discovery

Five years ago, „being found online“ was synonymous with „ranking on page one of Google.“ That equation no longer holds. Restaurant discovery has shifted to three new channels:

AI chatbots like ChatGPT (180+ million weekly users), Google Gemini (integrated into Google Search, billions of queries daily) and Perplexity have become the primary destination for restaurant searches with specific requirements.

AI-powered search in the form of Google AI Overviews now appears above classic search results and displaces blue links for intent-driven queries like „restaurant with allergy options.“

Voice assistants — Siri, Google Assistant, Alexa — answer millions of restaurant queries daily, based exclusively on Schema.org markup and Featured Snippets.

What all three channels have in common: they rely on structured, machine-readable data. Unstructured text, PDFs and Instagram photos provide zero input to these systems.

What AI Visibility Concretely Means

AI visibility is not a marketing term. It is a technical state: either your restaurant data exists in a format that AI systems can directly read and use — or it does not.

The four dimensions of AI visibility:

Dimension 1: Structural visibility. Are your dishes defined as individual, machine-readable entities? A dish is only AI-visible when it exists as a Schema.org MenuItem — with name, price, description, category, allergens and nutrition as separate data points. A menu as prose text or PDF does not meet this requirement.

Dimension 2: Allergen visibility. Specific allergen queries make up a large proportion of AI restaurant queries. „Gluten-free restaurant,“ „peanut-free dining,“ „lactose-free options“ — these queries can only be answered by AI when allergen data is machine-readable and structured per dish.

Dimension 3: Geographic and linguistic visibility. AI systems answer questions in the user’s native language. A restaurant available only in German does not exist for Chinese tourists, English-speaking expats or Turkish families in the AI — regardless of food quality.

Dimension 4: Trust signal visibility. AI systems source trust signals from controlled sources: Google Business Profile, Schema.org markup, llms.txt files and structured data feeds. Restaurants without these signals are bypassed for ambiguous queries.

→ Google AI Overviews for Restaurants

Why AI Ignores What Humans See

The underlying principle is simple but the consequences are serious: humans and AI „read“ websites in fundamentally different ways.

A human visits your menu page and immediately understands: „Ah, Kung Pao Chicken for €16.50, with peanuts, spicy.“ The association happens intuitively, from context and experience.

An AI reads the same page and sees: character strings without defined data types. It can guess that „16.50“ might be a price. It cannot say with certainty whether „peanuts“ is an allergen, an ingredient or a garnish element. It does not know whether the dish is currently available.

Now with Schema.org JSON-LD: The AI reads: object type „MenuItem,“ name „Kung Pao Chicken,“ price „16.50 EUR,“ allergen „peanuts (containsAllergen),“ category „Main Course,“ available „true.“ No interpretation required. One hundred percent precision.

This difference — interpretation vs. knowledge — determines whether your restaurant gets recommended for specific queries or not.

The Types of AI Queries Restaurants Lose

Without AI visibility, a restaurant systematically loses the following query types:

Allergen and diet queries are the most common lost queries: „Which Chinese restaurants offer gluten-free dishes?“, „Is there vegan dim sum near me?“, „Where can I eat safely with lactose intolerance?“ — Without machine-readable allergen data, your answer is: none at all.

Nutrition queries are growing: „Restaurant with meals under 600 calories,“ „high-protein dining in Frankfurt“ — only possible with NutritionInformation schema per dish.

Language-specific queries in English, Chinese, Turkish — consider: how many potential guests do not speak your local language as their mother tongue?

Agent-based bookings — in the future, AI agents will reserve tables on behalf of users. Without structured availability data and API access, restaurants will be bypassed for these autonomous bookings.

→ Google Business Profile and AI

→ Multilingual Menus and AI Visibility

How chiwai Creates AI Visibility

chiwai automates the entire AI visibility process in a continuous workflow:

Step 1 — Menu digitization: Upload supplier invoices, AI identifies ingredients and generates dish entries with all relevant metadata.

Step 2 — Schema.org generation: Every dish automatically receives complete JSON-LD markup — MenuItem, NutritionInformation, containsAllergen — machine-readable and directly processable by search engines and AI crawlers.

Step 3 — GBP push: Menu data is transmitted directly to Google Business Profile via the Food Menus API — no manual typing, no outdated data.

Step 4 — Multilingual: Automatic translation into DE/EN/ZH/ES/TR/AR and more languages on request ensures international AI queries find your restaurant.

Step 5 — llms.txt: The file is automatically created and kept current — AI crawlers receive immediate, precise information about your restaurant data.

The result: complete AI visibility across all relevant systems — Google Gemini, ChatGPT, Perplexity and voice assistants.

Start now — free plan available

FAQ

Why does ChatGPT recommend some restaurants and not others?

ChatGPT bases restaurant recommendations on data collected during its training crawls and retrieved in real-time via Bing when web search is enabled. It favors restaurants whose data is machine-readable and structured: Schema.org JSON-LD markup, complete Google Business Profile information and llms.txt files. Restaurants without these structured signals either do not appear in ChatGPT recommendations at all, or are only considered for very general queries („restaurant near me“). For specific queries — „peanut-free Chinese restaurant in Frankfurt“ — ChatGPT can only recommend restaurants where unambiguous allergen data exists as machine-readable facts. Restaurants without this data are simply bypassed, even if their food would actually be peanut-free.

Is AI visibility only relevant for large restaurant chains?

No — quite the opposite. Large restaurant chains like McDonald’s or Starbucks have dedicated technology teams managing structured data. For independent restaurants, AI visibility is the greatest opportunity to compete with chains on equal footing or even outperform them. A single Chinese restaurant with complete allergen schema, a 6-language menu and a maintained GBP profile is more visible for specific AI queries than a chain without this structuring. chiwai makes exactly this infrastructure accessible for individual restaurant operators — without an IT department, without technical knowledge.

How does AI visibility differ from local SEO?

Local SEO optimizes for traditional search engine rankings: Google My Business reviews, local keywords, NAP consistency (name, address, phone). AI visibility goes deeper: it optimizes the semantic quality and completeness of your data for AI systems. A restaurant can be excellently positioned for local SEO — page one for „Chinese restaurant Frankfurt“ — and still be invisible in ChatGPT and Gemini because no structured menu data exists. Ideally, the two combine: local SEO as the foundation, AI visibility as the next layer. With chiwai you build both layers simultaneously, because structured data improves both classical SEO and AI visibility.