Skip to content
Marketing, IT & AI automation · Surrey and Metro Vancouver
Digital Allo Marketing Services Ltd. Contact Us

Digital Allo Blog · November 19, 2026

How AI Search Decides Which Businesses to Recommend — and Why Your Reviews Decide It

By Digital Allo Editorial Team Published

A customer in Surrey used to Google “best roofer near me” and pick from a list. Now a growing share of them ask ChatGPT or Google’s AI instead — and the AI answers with one or two names, not forty. If your business is not one of the names, you never even get considered.

The question every Surrey business should be asking is no longer just “do I rank?” It is “what makes an AI assistant recommend me?” A research summary published 13 November 2026 pulls the 2026 studies together, and the answer is blunt: AI assistants recommend the businesses that other sources already vouch for. Your reviews, your directory listings and what other sites say about you matter more than anything on your own website.

Your customers have already moved

BrightLocal’s 2026 Local Consumer Review Survey found 45% of consumers used ChatGPT or other AI tools to find local businesses in 2026 — up from 6% a year earlier. That is the shift in one number: AI is no longer a novelty search for local services. It is where a large share of your next customers are starting.

And the stakes are high, because an AI answer does not behave like a results page. Ahrefs studied 300,000 keywords in February 2026 and found that when a Google AI Overview appears, the top-ranking page gets about 58% fewer clicks. Seer Interactive found pages cited inside an AI Overview earn about 2.1% click-through, against about 0.9% for uncited pages — and about 95% of comparison searches now show an AI Overview. Being named in the answer is worth far more than ranking tenth below it.

Reviews are the deciding factor

The clearest finding in the research is about reviews. Local businesses recommended by both ChatGPT and Perplexity averaged 133 Google reviews. Businesses invisible to AI averaged 11. That is not a small gap — it is the difference between being recommended and not existing in the AI’s world.

SOCi’s 2026 Local Visibility Index, covering more than 350,000 locations, found ChatGPT recommends only about 1.2% of local business locations. The bar is high, and it is set by third-party evidence: the businesses AI picks are the ones with a deep, recent review record, consistent listings and mentions elsewhere on the web.

This is the part Surrey businesses should take personally — including us. Digital Allo currently has 6 Google reviews. The businesses AI recommends average 133. Our competitors in the audit tell the same story: Shephard has 74, Elegant has 55. Reviews were already our number-one gap in our own SEO audit. This research says they are now the number-one gap for AI visibility too — and we are working on closing it the same way we tell clients to: by asking happy customers, every time.

Each AI engine is a different audience

One more reason not to optimise for a single tool: the engines barely agree with each other. A 2026 study of 602 prompts and 21,143 citations found ChatGPT cites about 6.9 sources per answer, Google’s AI Overviews and Gemini about 12.1, and Perplexity about 16.4. Another study found the overlap in cited domains between any two engines was at most 17%. Visibility has to be earned separately on each one.

Where they look also differs. Wikipedia and Reddit each account for roughly 12–13% of ChatGPT citations, with LinkedIn and YouTube also in the top ten. For agency recommendations, Clutch appears in about 84.5% of AI citations. The pattern is the same everywhere: the AI trusts what others say about you more than what you say about yourself.

What actually works on your own website

The original GEO study (KDD 2024) found that adding statistics, citing reliable sources and quoting experts raised visibility in generative engines by up to 40%. SE Ranking’s study of 129,000 domains found pages with 19 or more data points averaged 5.4 ChatGPT citations, against 2.8 for data-light pages. Comparison content made up 32.5% of citations across 4,100 B2B queries. In plain terms: pages that answer a question directly, back it up with real numbers from named sources, and honestly compare options get cited. Pages that just describe a service do not.

What does not work — save your money

Two popular “GEO quick fixes” have shown little direct effect, and Surrey businesses should not pay for them:

An llms.txt file. Across 137,000 domains, 97% of llms.txt files received no requests in a month, and Google says its AI features do not need one.

Extra schema markup for AI. Ahrefs tested adding schema to 1,885 already-cited pages in May 2026 and citations barely changed. Schema still helps search engines understand a page, so keep what you have — but buying a “GEO schema package” to win AI citations is paying for something the evidence does not support.

Google’s own May 2026 guidance says the same thing in plainer language: optimising for AI features is still SEO, no special optimisation is needed, and it warns against seeking inauthentic mentions. Buying fake reviews or fake mentions is the fastest way to be filtered out of the answers you want to appear in.

One more thing: don’t misread your Search Console numbers

Google confirmed on 13 November that a logging bug under-reported impressions in the Search Console Generative AI performance report for 9–11 November. It affected data logging only, not rankings — but the November core update started on 10 November, inside the same window, so a dip in that report and a ranking move are easy to confuse. Check the dashboard, not just the chart. The core update is still rolling out; judge nothing until it finishes and a full week has passed.

What to do this week

  • Ask three happy customers for a Google review — and reply to every review you get. Reviews are now the single biggest AI-visibility signal a local business controls.
  • Check your listings match everywhere — same name, address, phone, services and prices on your website, your Google Business Profile and your directories. AI answers are built from what it finds about you.
  • List your business where AI actually looks — the two or three directories your industry’s AI citations come from, not twenty random ones.
  • Open key pages with a one-sentence answer, then prove it — real numbers, named sources, expert quotes. Pages with data get cited; pages with adjectives do not.
  • Publish one honest comparison — your service against the alternatives, with real differences. Comparison content earns a disproportionate share of citations.
  • Test the same customer questions in each AI engine monthly — ask ChatGPT, Google AI Mode, Perplexity and Copilot the questions your buyers actually ask, and track which firms and sources come up.
  • Don’t buy llms.txt or “GEO schema” packages — spend that money on reviews and on pages worth citing instead.

The search page is becoming an answer page, and the answer names one or two businesses. The businesses that get named are not the ones with the cleverest tricks. They are the ones other people already vouch for — in reviews, in listings, and in pages that prove what they claim.

Call Digital Allo at (604) 401-0699 or visit https://www.digitalallo.com/contact-us/ — 13898 64 Ave #108, Surrey, BC.

A note on the numbers in this post: every figure comes from a third-party study cited in Allway Media Design’s 13 November 2026 research summary, with the original study named each time. Several are vendor research and should be read as directional, not exact — the pattern across them is what matters.

FAQ

How does AI search decide which businesses to recommend? AI assistants recommend the businesses that other sources already vouch for. Third-party signals — reviews, directory listings, media coverage and community discussion — carry the most weight, followed by clear, specific, well-sourced pages on your own site.

How many Google reviews does a business need to be recommended by AI? There is no fixed threshold, but the gap in the research is stark: local businesses recommended by both ChatGPT and Perplexity averaged 133 Google reviews, while businesses invisible to AI averaged 11. More real, recent reviews — and replies to them — is the single biggest lever a local business controls.

Why do ChatGPT, Google and Perplexity recommend different businesses? They cite different sources: ChatGPT averages about 6.9 sources per answer, Google’s AI Overviews about 12.1 and Perplexity about 16.4, and the overlap in cited domains between any two engines is at most 17%. Visibility has to be earned separately on each one.

Does an llms.txt file help a business appear in AI answers? No, based on current evidence. Across 137,000 domains, 97% of llms.txt files received no requests in a month, and Google says its AI features do not need one.

Does adding schema markup win AI citations? Not on its own. Ahrefs found adding schema to 1,885 already-cited pages barely changed AI citations. Schema still helps search engines understand a page, but it is not a shortcut to being cited by AI assistants.

Should a Surrey business worry about the November 2026 core update and AI visibility at the same time? The core update is still rolling out, so don’t judge any ranking move until it finishes and a full week has passed. The AI-visibility work — reviews, consistent listings, pages with real data — is the same foundation either way, so it is worth doing now regardless of the update.

Service areas

Serving businesses across British Columbia

On site and remote across Metro Vancouver and the Fraser Valley. Remote-first on Vancouver Island and in the Okanagan, with visits by arrangement.

Surrey & South of the Fraser

On-site and remote

Vancouver, Richmond & North Shore

On-site and remote

Burnaby, Tri-Cities & Maple Ridge

On-site and remote

Fraser Valley

On-site and remote, with planned visits further east

Vancouver Island

Remote-first, visits by arrangement

Okanagan & Thompson

Remote-first, visits by arrangement

See all service areas