Opening hours, dogs, the terrace: why conflicting data drops your café from AI answers

Why AI recommends your café — or skips it entirely

Today, someone looking for somewhere nearby to have breakfast with a dog increasingly doesn't type that into Google. They ask ChatGPT, Google Gemini or read the AI Overview sitting above the regular results. None of these tools opens your nicely designed website first. They stitch an answer together from many sources at once — your Google Business Profile, your site, delivery and menu portals, reviews, local directories — and present it as one confident answer.

The problem: these models are built to sound certain. When your data matches everywhere, your café reads as a safe recommendation. When it contradicts itself, the model notices the inconsistency, can't resolve it, and quietly picks the café down the street whose facts line up instead. You're not being penalized for being bad. You're being skipped for being unclear.

For a café that's a rough deal, because the exact questions that decide a visit — what time do you open, can the dog come, is there a terrace, is there vegan cake — are also the ones cafés are most likely to answer inconsistently across the web. Getting those right is what decides whether the AI names you at all.

Opening hours: the classic mistake that trips up most cafés

Opening hours cause more confusion than anything else. A typical café lists them in at least four places: the Google profile, the website footer, Facebook, and a delivery or reservation platform. Then the terrace season shifts the summer schedule, Monday becomes a closing day, or Sunday opening moves from 8 to 9. You update Google and forget the other three. Now there are four different versions of the truth online.

For an AI, that's a red flag. Ask "Is Café Löffelweise open on Sundays?" and the model finds "from 8," "from 9," and "closed Sundays" all at once. It has no way to know which one is current. Its safe answer becomes "hours vary, check directly" — or it names a different café altogether. Either way, you lose the guest who wanted breakfast right now.

The fix isn't glamorous, but it works: keep one master list of your hours, including holidays and seasonal changes, and update every platform from that single list every time something changes. Holiday hours in your Google profile matter especially, since that's exactly what people ask AI about most.

Dogs welcome? The question almost nobody answers in writing

"Is this café dog-friendly?" is one of the questions people ask AI most often about cafés — and one that almost no café website answers directly. You know dogs are welcome. There's a water bowl on the terrace. But is any of that written down anywhere? Usually not. And what isn't written down doesn't exist for an AI.

It gets worse when the signals conflict. One review says "great that dogs are allowed," another says "they turned us away with our dog" — maybe because the rule differs indoors versus outdoors. Your website says nothing either way. The AI now has a yes, a no, and silence from the one source that should settle it. The result is a hedge that discourages guests instead of drawing them in.

Be explicit. Put a clear sentence on your website and in your Google profile: "Dogs are always welcome indoors and on the terrace, and we keep a water bowl out for them." Turn on the "dog-friendly" attribute in your Google profile if it's available. One unambiguous sentence outweighs ten conflicting reviews.

Terrace, seating, indoors vs. outdoors: keeping your space facts consistent

The terrace is often a café's best feature — and also a data mess. Instagram shows it full in summer; by winter it's been taken down. Ask in March whether the café has a terrace and the honest answer is "yes, from April." If that's written nowhere, the AI has to guess. It might say yes when the terrace is currently closed, or "unclear" when you actually have one of the nicest terraces in town.

Seating gets misrepresented too. A twelve-seat café marketed as a party venue confuses guests and AI alike. Ask "can eight of us sit together?" and the answer needs to come straight from your data. If your website says "cozy nooks" while a booking portal claims "ideal for groups up to 20," that contradiction costs you reservation inquiries.

Describe your space plainly: how many indoor and outdoor seats, when the terrace is open, whether it's covered, whether it's heated. These details sound minor, but they're exactly what AI answers are built from. Concrete numbers beat marketing adjectives every time.

Menu and dietary options: vegan, gluten-free, lactose-free

"Do they have vegan cake?" or "any gluten-free options?" are make-or-break questions, especially for people managing an intolerance who need certainty before they show up. If your menu only exists as a nicely designed PDF or an Instagram photo, AI systems often can't read it reliably. A photo of your chalkboard looks great to a person and is frequently invisible to a machine.

Here's the classic contradiction: your website promises "fresh vegan cake, every day," but the daily selection actually varies and some days there isn't one. Then a review complains "no vegan cake here." Now the AI has its least favorite situation — a claim and a counterexample side by side. It hedges, saying something like "sometimes has vegan options" instead of a clean yes.

Put your core offerings into real, readable text — not just an image. Word it honestly, in a way that stays true every day: "We always have at least one vegan and one gluten-free cake on offer." If you say it, deliver on it. When your claims match reality, reviews back you up instead of contradicting you.

Reviews: the AI signal you can actually shape

AI models pay close attention to reviews, because reviews are written in real guest language. When enough reviews mention "quiet café, good for working, solid WiFi," "best flat white in town," or "our dog always gets a warm welcome," the model picks up your strengths straight from them — and those phrases start showing up almost verbatim in its recommendations. Reviews aren't just a star rating. They're training material for your AI visibility.

That's why it pays to ask happy guests for reviews with real detail, not scripted ones. A regular who writes "finally a café with a big terrace, vegan cake, and my dog is welcome too" hands you three confirmed facts in a single sentence — worth far more than ten reviews that just say "nice."

Reply to reviews too, including the critical ones. If someone writes "I thought dogs were allowed" and you answer "dogs are always welcome on our terrace — indoors we ask owners to keep them close, out of consideration for other guests," you've corrected the record in public. The AI reads your reply and picks up the actual rule.

One source of truth: your NAP and core data

At its core this comes down to one idea: a single truth, repeated identically everywhere. The industry calls the basics NAP — name, address, phone number. For a café, the list is longer: hours, terrace, dog policy, dietary options, WiFi, whether you take reservations. These core facts need to match across every platform. Even a misspelled café name or an outdated phone number can throw off how an AI matches your listing to the rest of your data.

Do a full, honest audit once. Open your Google profile, your website, Facebook, Instagram, every delivery portal, and every directory you're listed in. Put the core facts into a table and mark every mismatch. You'll almost certainly find several you didn't know about — stale hours, a wrong phone number, a three-year-old listing nobody updated.

Then fix them, one by one. It's tedious work, but it's the single highest-leverage thing you can do for AI visibility. A café with boring, consistent data gets recommended more reliably than the trendiest café in town with facts that don't add up.

A 7-point plan for AI-visible café data

None of this requires technical skill — it takes discipline more than expertise. Set these up properly once, then maintain them every time something changes, and you'll stand out from most cafés that leave their data to chance.

Order matters here: clear up existing contradictions first, then build clear, positive statements on top of a clean base. And everything you write down has to match what actually happens in the café — otherwise your guests' reviews will start contradicting you all over again.

• Keep your hours identical everywhere — holidays and terrace season included — and treat one master list as the single source.
• State your dog policy in one clear sentence on your website and Google profile, spelling out indoors and outdoors separately.
• Describe the terrace concretely: how many outdoor seats, what season it's open, whether it's covered, whether it's heated.
• Publish your menu as readable text, not only as an image or PDF, and call out vegan and gluten-free options by name.
• Ask guests directly for honest, detailed reviews, and reply to every one you get.
• Run a NAP audit: compare every platform, list the mismatches, and fix them.
• Recheck everything every few months so new contradictions don't creep back in.