Original research · 2026-07 edition

AI SEO Statistics: Cafe (2026-07 edition)

15 questions · 45 AI responses · 3 models · measured 2026-07-04

See how AI talks about YOUR businessFull data table

The question bank

The questions we tested — sampled from real buyer journeys in cafe.

Each model answered every question once, same wording, same day. These are the prompts behind every percentage on this page.

Where is a good spot to host a casual 1-on-1 business meeting that isn't too noisy or crowded?
Is it actually worth paying extra for a cafe that roasts their own beans on-site?
How do I find a coffee shop that is actually laptop-friendly and has enough power outlets for a long session?
What are the tell-tale signs of a high-quality espresso bar versus a generic fast-food coffee place?
I want to organize a small weekend meetup; how many shops should I call to check for group seating?
Are there specific things I should ask a barista to see if they really know their craft and bean origins?
Is it considered rude to sit in a cafe for three hours if I only bought one medium drip coffee?
What is the average price for a bag of local craft coffee beans compared to standard grocery store brands?
Show all 15 questions
I need a cafe with a patio that allows dogs but also provides decent shade and Wi-Fi for working.
Why do some modern cafes charge a service fee on top of the tip and is that becoming the industry standard?
What is the best way to judge a cafe's food menu quality before I actually commit to going there?
I am looking for a place that serves authentic ceremonial grade matcha rather than just sugary pre-mixed lattes.
Should I trust online reviews that complain about slow service at a specialty coffee shop or is that just part of the process?
What are the red flags that a cafe is using low-quality or expired syrups and milk alternatives in their drinks?
How can I find a local cafe that regularly hosts community events like open mic nights or book clubs?

Model by model

17-point average divergence: which AI you ask changes the answer.

The divergence index is the average gap between the most and least likely model per behavior. Higher = the models disagree more about cafe buyers.

Behavior rates across 15 cafe buyer questions, 2026-07 edition. Last column: average across models.
ChatGPTClaudeGeminiConsensus
Recommends hiring a professional7%7%7%100%
Suggests DIY first20%7%20%80%
Names specific providers13%13%33%67%
Gives price or cost info13%13%20%80%
Tells to check reviews40%20%13%53%
Tells to verify credentials13%0%0%87%
Mentions case studies / portfolio0%0%0%100%
Mentions local proximity53%27%33%53%
Gives selection criteria73%60%47%60%
Warns about red flags40%33%27%80%
Asks a clarifying question67%40%13%40%
Recommends multiple quotes7%7%0%87%

By model

How each assistant handled Cafe questions.

Reading the 45 answers model by model shows how differently the three assistants treat the same cafe questions. On the most consequential behavior — whether to send the buyer to a professional at all — the rate ranged from 6.7% (ChatGPT) down to 6.7% (ChatGPT), a 0-point gap on an identical question set.

Across the 15 cafe answers it produced, ChatGPT recommended hiring a professional in 6.7% of them and suggested a DIY approach first 20% of the time. It named a specific provider in 13.3% of answers (about 0.4 distinct providers per answer) and included price or cost information 13.3% of the time. ChatGPT asked a clarifying question before answering in 66.7% of cases, warned about red flags or scams in 40%, and told the buyer to verify credentials in 13.3%, averaging 420 words per answer. On the remaining cues it told the buyer to check reviews in 40%, pointed to case studies or a portfolio in 0%, and framed the choice around local proximity in 53.3%; a selection-criteria checklist appeared in 73.3% of its answers and a recommendation to gather multiple quotes in 6.7%.

Across the 15 cafe answers it produced, Claude recommended hiring a professional in 6.7% of them and suggested a DIY approach first 6.7% of the time. It named a specific provider in 13.3% of answers (about 0.1 distinct providers per answer) and included price or cost information 13.3% of the time. Claude asked a clarifying question before answering in 40% of cases, warned about red flags or scams in 33.3%, and told the buyer to verify credentials in 0%, averaging 266 words per answer. On the remaining cues it told the buyer to check reviews in 20%, pointed to case studies or a portfolio in 0%, and framed the choice around local proximity in 26.7%; a selection-criteria checklist appeared in 60% of its answers and a recommendation to gather multiple quotes in 6.7%.

Across the 15 cafe answers it produced, Gemini recommended hiring a professional in 6.7% of them and suggested a DIY approach first 20% of the time. It named a specific provider in 33.3% of answers (about 0.9 distinct providers per answer) and included price or cost information 20% of the time. Gemini asked a clarifying question before answering in 13.3% of cases, warned about red flags or scams in 26.7%, and told the buyer to verify credentials in 0%, averaging 225 words per answer. On the remaining cues it told the buyer to check reviews in 13.3%, pointed to case studies or a portfolio in 0%, and framed the choice around local proximity in 33.3%; a selection-criteria checklist appeared in 46.7% of its answers and a recommendation to gather multiple quotes in 0%.

Taken together, ChatGPT is the assistant most likely to route a cafe buyer to a professional (6.7%) and ChatGPT the least (6.7%). ChatGPT produced the longest answers, at 420 words on average. Specific providers were named most often by Gemini (33.3%) — even there, roughly one answer in 3 carried a name.

Where they disagree

The behaviors where the choice of model changes the answer.

The divergence index for this study is 17.4 points — the average distance between the most and least likely model across the coded behaviors. The gaps below are where which assistant a cafe buyer happens to ask matters most:

  • Asks a clarifying question: from 13.3% (Gemini) to 66.7% (ChatGPT) — a 53-point spread.
  • Tells the buyer to check reviews: from 13.3% (Gemini) to 40% (ChatGPT) — a 27-point spread.
  • Mentions local proximity: from 26.7% (Claude) to 53.3% (ChatGPT) — a 27-point spread.
  • Gives selection criteria: from 46.7% (Gemini) to 73.3% (ChatGPT) — a 27-point spread.
  • Names a specific provider: from 13.3% (ChatGPT) to 33.3% (Gemini) — a 20-point spread.

The widest single gap — asks a clarifying question, 53 points — means a cafe buyer can receive materially different guidance on the same question depending only on which assistant they happen to open, so any visibility strategy built on a single model's behavior describes only part of the cafe market.

Where they agree

The points of near-consensus in Cafe.

On other behaviors the three models move almost in lockstep — the points of near-consensus for cafe, where all three landed within a few points of each other:

  • Recommends hiring a professional: 6.7% across all three models.
  • Mentions case studies or portfolio: 0% across all three models.
  • Gives price or cost information: 13.3%–20% across all three (a 7-point spread).
  • Recommends multiple quotes: 0%–6.7% across all three (a 7-point spread).

Measured question by question, the three assistants coded a response the same way most consistently on "recommends hiring a professional" (identical coding in 100% of questions) and least consistently on "asks a clarifying question" (40%).

Every behavior, measured

All twelve coded behaviors for Cafe, averaged across the three models.

The behaviors AI models reproduce most often for cafe are gives selection criteria (60% on average), asks a clarifying question (40%) and mentions local proximity (37.8%); the rarest are mentions case studies or portfolio (0%), tells the buyer to verify credentials (4.4%) and recommends multiple quotes (4.5%). Each figure below is the share of a model's 15 answers in which the behavior appeared at least once, averaged across the 3 models with the full per-model range in parentheses:

  • Gives selection criteria: 60% on average (ChatGPT 73.3%, Claude 60%, Gemini 46.7%) — a 27-point spread.
  • Asks a clarifying question: 40% on average (ChatGPT 66.7%, Claude 40%, Gemini 13.3%) — a 53-point spread.
  • Mentions local proximity: 37.8% on average (ChatGPT 53.3%, Claude 26.7%, Gemini 33.3%) — a 27-point spread.
  • Warns about red flags or scams: 33.3% on average (ChatGPT 40%, Claude 33.3%, Gemini 26.7%) — a 13-point spread.
  • Tells the buyer to check reviews: 24.4% on average (ChatGPT 40%, Claude 20%, Gemini 13.3%) — a 27-point spread.
  • Names a specific provider: 20% on average (ChatGPT 13.3%, Claude 13.3%, Gemini 33.3%) — a 20-point spread.
  • Suggests a DIY approach first: 15.6% on average (ChatGPT 20%, Claude 6.7%, Gemini 20%) — a 13-point spread.
  • Gives price or cost information: 15.5% on average (ChatGPT 13.3%, Claude 13.3%, Gemini 20%) — a 7-point spread.
  • Recommends hiring a professional: 6.7% on average (ChatGPT 6.7%, Claude 6.7%, Gemini 6.7%).
  • Recommends multiple quotes: 4.5% on average (ChatGPT 6.7%, Claude 6.7%, Gemini 0%) — a 7-point spread.
  • Tells the buyer to verify credentials: 4.4% on average (ChatGPT 13.3%, Claude 0%, Gemini 0%) — a 13-point spread.
  • Mentions case studies or portfolio: 0% on average (ChatGPT 0%, Claude 0%, Gemini 0%).

Trust signals

How well the models protect the cafe buyer.

Beyond whether to hire, the rubric codes how carefully each assistant protects the cafe buyer once a decision is made. Telling the buyer to check reviews or ratings appeared in 24.4% of answers on average. Verifying credentials or certifications appeared in 4.4%. Warning about red flags or scams appeared in 33.3%.

On structuring the decision, a selection-criteria checklist showed up in 60% of answers on average and a recommendation to gather multiple quotes in 4.5%. The single least-reproduced protective signal for cafe is "tells the buyer to verify credentials" at 4.4% on average — the clearest opening for content that supplies it, since the models are not yet reliably surfacing that guidance on their own.

Referral behavior

Do AI models name Cafe providers?

For service providers the decisive question is whether these systems name anyone at all. Across 45 cafe answers, a specific provider was named in 20% of responses on average — roughly 0.5 distinct providers per answer. In practice the assistants behave far more as an explanatory layer than as a referral engine for cafe: visibility comes from being the reasoning a model reproduces, not from being the named recommendation.

The question set

What these 15 Cafe questions cover.

The 15 questions behind every percentage on this page were drawn from real cafe (hospitality; buyer hiring decisions for this specific service) buyer journeys. Each was put to all 3 models once, with identical wording, so the rates above describe how the assistants handled this exact cafe question set — not a general prior or a hand-picked subset. The full list is shown earlier on this page; the coded percentages are what those specific questions produced.

How to read this

A note on the numbers.

A percentage here is the share of a model's 15 answers in which the behavior appeared at least once — not a confidence score. Because each model answered every question exactly once on 2026-07-04, the figures describe this specific cafe question set and snapshot rather than a general prior. The full protocol and coding rubric are documented in the study methodology.

Methodology

A controlled snapshot, documented end to end.

15 standardized buyer questions per industry, one response per model per question (ChatGPT (gpt-5-mini), Claude (claude-sonnet-5), Gemini (gemini-3-flash-preview)), collected 2026-07-04, coded against a fixed 12-behavior rubric with human QA. AI outputs vary with model version, location and time — figures describe this sample and window, and are refreshed each edition. Read the full methodology →