AEO for Calgary Restaurants.

Before you optimize anything, understand the queries. The dining-related prompts AI engines see most often in Calgary, based on our own usage and on what generates AI Overviews when we test:

• Neighborhood + cuisine. “Best ramen in Calgary,” “where to eat in Inglewood,” “Italian restaurants near Stephen Avenue.”
• Occasion-driven. “Best date night restaurants Calgary 2026,” “where to take a client for dinner downtown Calgary,” “anniversary dinner Calgary.”
• Constraint-driven. “Vegan restaurants Calgary,” “gluten-free brunch Calgary,” “Calgary restaurants open late,” “kid-friendly restaurants Calgary.”
• Comparison-driven. “Is X restaurant or Y restaurant better for date night,” “Model Milk vs Bridgette Bar.”
• Decision-stage. “What should I order at <restaurant>,” “is <restaurant> good for groups.”

Note what is missing from this list: brand-name lookups. Diners who already know your restaurant name will Google it directly. AI engines win the discovery query, the one where the diner has a need but no specific name in mind. Every one of those queries is a chance to be the answer or to be invisible.

These are practical parts of a crawlable, verifiable restaurant site. Retrieval and source-selection systems change, and none of the five guarantees a citation.

1. Restaurant or LocalBusiness JSON-LD with depth

Schema.org markup states what your business is in machine-readable form. Google says structured data is not required for its generative AI search features, but it still helps search engines read your facts. The restaurant-grade version uses the Restaurant type (a subclass of LocalBusiness) and includes name, address, geo, servesCuisine, priceRange, openingHours, menu, acceptsReservations, telephone, and url. In our view, filling in every field that applies is the most useful technical fix on this list.

2. llms.txt at the site root

An llms.txt file is an optional Markdown summary at yourrestaurant.com/llms.txt. Compatible tools may use it to discover selected public pages, but support is incomplete and Google does not use it. Read the full llms.txt reference before deciding whether to publish one.

3. Real content depth, not a menu PDF

Crawlable prose about the restaurant, chef, cuisine, and menu gives retrieval systems more factual material than a menu PDF or one-line homepage. There is no universal word-count floor. Write enough to answer real diner questions without padding the page.

4. Named author or chef E-E-A-T

Google describes E-E-A-T (Experience, Expertise, Authoritativeness, Trustworthiness) as what its systems look for in helpful content; no AI engine publishes an equivalent. The cheapest fix: add a chef bio page with a Person JSON-LD block, link to it from the homepage, and put a <meta name="author"> tag in the HTML head. In our Calgary audit, only 7% of small business sites had any named author signal. For a restaurant, putting a real chef and their story on the site is also great brand work, so this is not just a technical fix.

5. Third-party mentions

In our view, it helps when your restaurant is mentioned elsewhere: Reddit threads about Calgary dining, local blog roundups, news features in Avenue Magazine and Calgary Herald, Eater Calgary if you can land it, even Yelp and OpenTable reviews. You cannot fake this overnight, but you can earn it with PR outreach, by pitching local food writers, by showing up in roundups about your neighborhood, and by being genuinely good. Real third-party coverage gives AI engines more independent sources to check, though no engine publishes how it weights them.

If you do nothing else, do these seven things in this order. Each one is well-scoped and finishable in an afternoon.

1. Add Restaurant JSON-LD with all 11 fields (name, address, geo, servesCuisine, priceRange, openingHours, menu URL, acceptsReservations, telephone, url, image). Validate at search.google.com/test/rich-results before shipping.
2. Consider an optional llms.txt file. Publish one only if a short, maintained summary of your public pages is useful. It does not guarantee citations.
3. Check retrieval access in robots.txt. Confirm that OAI-SearchBot, ChatGPT-User, Claude-SearchBot, Claude-User, and PerplexityBot are not blocked. Decide training crawler access separately.
4. Write a real About or story page. Explain who you are, why the restaurant exists, the cuisine philosophy, and the chef’s background. There is no universal word-count floor.
5. Add a chef bio with Person schema. Include a real photo, relevant experience, and an accurate JSON-LD Person block. Write enough to establish who the chef is without padding the page.
6. Publish your menu as crawlable HTML, not a PDF. Some AI crawlers parse PDFs poorly, and a PDF menu is harder to keep current. Put it in actual page markup with section headings, item names, and a one-line description per item. PDFs can stay as a download link, but the canonical menu must be HTML.
7. Add FAQ schema only to a real visible FAQ section. Cover the questions diners actually ask: parking, reservations, dietary accommodations, group sizing, dress code, and average meal duration. The visible answers and JSON-LD must match exactly. FAQ parity is an optional 2-point audit check, not a ranking or citation lever.

The published study did not break results out by restaurant, so these are the pass rates across all 123 audited Calgary small business sites, restaurants included:

Translation: almost every site lets AI crawlers in, but few ship the structured signals. For a restaurant, a PDF-only menu and a missing story page widen that gap.

Imagine two Italian restaurants in Inglewood with similar food, pricing, and reviews. Restaurant A has a hero image, a menu PDF, a map, and a reservations link. Restaurant B adds accurate Restaurant JSON-LD, a chef bio, a crawlable HTML menu, and a visible FAQ section with matching schema. It may also publish an optional llms.txt summary. The stronger site gives retrieval systems more public facts to verify, but no single addition guarantees a citation.

When a diner asks Perplexity for the best Italian restaurant in Inglewood for date night, Restaurant B offers more verifiable source material: a cuisine description, named chef, reservation policy, and crawlable menu. That can improve retrieval, but the source-selection result still depends on the engine and the wider evidence available.

This is the gap that exists across nearly the entire Calgary restaurant scene right now. Closing it is structural, not creative, and the structural fixes are well-known.

From there, improve named chef or author information and add FAQPage schema only where visible questions and answers already exist. Those changes make the page easier to interpret, but they do not guarantee rankings or citations.

Restaurants that improve crawlability, structured data, authorship, and source quality give retrieval systems better material to work with. That still does not guarantee citations, so track a stable set of Calgary dining prompts over time. AI answers are one channel, not the whole plan, so pair this with the rest of your Calgary restaurant marketing.

Related: Restaurant website design, the menu is the website.

• SaaS AEO: How B2B Tools Get Recommended by ChatGPT in 2026