Key takeaways
- Google ranks whole pages for a position in a list. An AI assistant assembles one answer and credits a few sources, so it rewards the clearest quotable passage.
- The overlap is smaller than most people assume, though every public number comes from vendors selling AI search tooling.
- Both fights stand on the same foundation: a fast site with real HTML, one job per page, consistent details, and real reviews.
- Google documents that AI Overviews need no special optimization, because they run on the core Search ranking systems.
- Short local searches still resolve in the map pack. Long full-sentence questions are where AI answers decide who gets named.
On this page
Are Google ranking and AI visibility the same problem?
No. Ranking on Google is a contest for a position in a list of links, judged page against page. Getting cited in an AI answer is a contest to be the source a model quotes while it writes a single answer from several pages. The foundation is shared, the finishing work differs, and a business can genuinely win one while losing the other.
The reason this matters to your wallet is that the industry has quietly merged the two into one product. AI SEO packages, AI visibility retainers, generative engine optimisation. Some of that work is real. A lot of it is the same technical and content fundamentals you were already buying, resold at a premium because the word AI is in the title.
Separating the two lets you ask a sharper question: which one is actually costing me customers this quarter. For most local businesses I look at, the answer is neither. It is that the site is slow, the hours are wrong on three directories, and nobody has written down what the business charges. That is the boring foundation both fights are built on.
Same footing, different finishing. Google decides which page deserves the click. An assistant decides which sentence deserves the credit.
How much do Google results and AI citations actually overlap?
Less than you would expect, going by the numbers in circulation, and every one of those numbers comes from a company selling AI search tooling. The honest position is that the overlap has clearly shrunk, the exact figure is contested, and the direction matches what you see when you run the test on your own queries.
The two figures quoted most often: one 2026 analysis published by SEO.com found the overlap between Google top ten results and what AI engines cite has fallen below 20 percent, and a related write-up from Goodfellas Tech reports that roughly 90 percent of ChatGPT citations come from pages outside Google top 20. I quote both with their sources attached because that is the only responsible way to use a vendor-published statistic.
You can check the shape of the claim yourself in ten minutes, which is what I would rather you do. Take one query your customers use. Look at the first page of Google. Then ask ChatGPT and Perplexity the same thing as a full sentence and note which sites get cited. The gap you see with your own eyes is worth more to your decisions than any percentage on a vendor blog.
What do both fights reward?
The same unglamorous foundation, which is the good news, because it means most of the work counts twice. A fast site that ships real HTML, one clear job per page, details that agree everywhere you appear, real reviews, and genuine expertise written down. Get this right and you are competitive in both contests before you optimise anything.
- A site that loads as real HTML, quickly, on a phone, because both a crawler and a customer give up at the same point.
- One clear job per page, with a heading that names it and content that delivers it.
- Details that agree everywhere: hours, service area, phone number, and prices identical on the site, the Google Business Profile, and every directory.
- Real reviews from real customers, since volume and rating are a documented local ranking input and a trust signal a model can read.
- Genuine expertise written down in your own words, with the specifics only your business could state.
- Internal links with descriptive anchors, so both a crawler and a reader can follow the trail from one page to the related one.
Google is unusually blunt about part of this. Its guidance on AI features says there are no additional requirements to appear in AI Overviews or AI Mode and no special optimizations necessary, because the generative features run on the core Search ranking systems. For that surface, ordinary search fundamentals are the documented strategy, full stop.
If you are earlier in the journey and still deciding whether search work is worth anything yet, do I need SEO right away answers the prior question honestly, and the local search visibility guide covers who is actually looking for you nearby.
Where do the two fights genuinely diverge?
Four places: the unit that wins, what you get when you win, how much format matters, and how much corroboration matters. None of these are exotic. They just explain why a page can sit at position three on Google and never once be named by an assistant answering the same question.
- 01
The unit that wins is different
Google ranks pages against each other for a position in a list. An assistant assembles one answer from several sources and credits a few. So Google rewards the strongest page overall, while an assistant rewards the clearest passage inside any page it happened to open.
- 02
The click is not the prize
A top Google result earns a visit. A citation in an AI answer sometimes earns a visit and sometimes only earns your name being said out loud to a customer who then calls you directly. Both are worth having, and only one of them shows up cleanly in your analytics.
- 03
Format carries more weight
A model summarising four pages will lean on the one with a clean comparison table, a numbered process, or an answer sitting in the first forty words. Ranking systems care about that less. Structure is a tiebreaker on Google and closer to a deciding factor in an AI answer.
- 04
Corroboration matters more
Assistants hedge toward things they can see confirmed in more than one place. A price stated on your site and matching a directory listing and your own pricing file is safer to repeat than a number that appears once. Google will index a single claim happily.
| Ranking on Google | Being cited in an AI answer | |
|---|---|---|
| What competes | Whole pages against each other | Passages inside whichever pages get retrieved |
| What you win | A position, and the click that follows | Your name in the answer, sometimes with a link |
| Biggest lever | Relevance, reputation, and the local pack for local queries | A self-contained answer in the first forty words |
| Format weight | A tiebreaker | Close to a deciding factor |
| Corroboration | A single clear claim indexes fine | Facts confirmed in several places get repeated |
| Measurement | Search Console, positions and clicks | Manual testing, monthly, written down |
Read the table as one story. Everything on the right is achievable by rewriting pages you already have. That is genuinely encouraging: the AI half of this is mostly an editing job, and the editing makes the pages better for the humans skimming them on a phone in a truck.
What happens to the click when an AI answers first?
Some of it disappears, and the size of the loss is not well established outside the US. The one credible study is Pew, which tracked 900 US adults across 68,879 searches and found link clicks fell from 15 percent to 8 percent when an AI summary appeared. No Canadian sample and no local-intent breakout, so it tells you the direction and not your number.
For a local service business, this stings less than the headline suggests. A customer looking for a furnace repair in Trail was never going to read three blog posts. They wanted a name and a phone number, and they have been getting that from the map pack for years without opening a single website. The AI answer is doing the same job in a different wrapper.
Where it does change things is the research stage. People ask assistants what something should cost, what to look out for, and whether they need a professional at all. If your site answers those questions honestly, your business gets named during the part of the process you never used to see. That is also why publishing real prices or ranges is worth the discomfort, a case I make in should a professional firm publish its fees.
Does any of this change things for a business in a small town?
Partly. Short local searches still resolve in the Google map pack, where your Business Profile, your categories, your distance from the searcher, and your reviews decide the outcome. Full-sentence questions are the ones landing in AI answers, and that is a growing share of how people research before they call.
- Short local queries
- Plumber in Castlegar, dentist near me, coffee Nelson. These still resolve in the Google map pack, where categories, distance, reviews, and profile completeness decide the outcome. AI answers barely enter the picture.
- Long question queries
- Who does emergency furnace repair in the West Kootenay on a weekend, and roughly what does a call-out cost. This is the shape people now type into an assistant, and the businesses with clear written answers are the ones that get named.
- Research before the call
- Comparison and how-much questions asked days before anyone books. Assistants are heavily used here, which means your pricing honesty and your explanations are doing sales work while you sleep.
- The honest gap
- Nobody has credible Canadian local data on any of this yet. Every study in circulation is either US-only or published by a company selling AI search tooling. Anyone quoting you a precise Canadian local number is guessing with confidence.
Google names relevance, distance, and popularity as the local ranking factors, and states directly that the categories you select affect your local ranking. That single field is worth more to a Castlegar or Nelson business than every AI tactic in circulation, and it takes ten minutes. Reviews matter too: volume and rating are a documented input, and I would reply to them anyway for the humans reading, which is the case made in reviews, local search, and the calls they bring.
The Kootenay angle is real and it has a release valve. Small-town customers can smell a fake answer in one sentence, which is exactly why plain honest pages win here. The same pages work identically for a client three provinces away, because clarity does not have a postal code.
Where should a small budget go first?
Into the shared foundation, every time, because it is the only work that pays into both fights at once. Speed, real HTML, one job per page, consistent details, a complete Business Profile, and answers written in the first paragraph. Only after that should you chase whichever gap your own monthly testing reveals.
- 1Fix the shared foundation first: load speed, real HTML, one job per page, and details that agree everywhere. It is the only work that pays into both fights at once.
- 2Complete your Google Business Profile properly, with the fewest accurate categories, since Google names categories as the one profile field it ties to local ranking.
- 3Rewrite the opening paragraph of your three most important pages so each answers its own question in the first forty words.
- 4Turn your five most-asked customer questions into a visible questions and answers block, phrased the way customers phrase them.
- 5Publish the one page only you could write: your real process, your real prices or ranges, and what you will not take on.
- 6Test monthly. Search your key queries on Google, then ask the same thing in ChatGPT and Perplexity, and write down who won each.
- 7Spend on the gap the test reveals, one fix at a time, and stay away from any package that promises to solve both at once.
What I would avoid is buying a package that promises both fights solved, priced by the month, with a dashboard attached. The measurement to justify it does not exist yet: AI Overviews traffic is folded into the regular Web search type in Search Console, with no separate breakout. If a report claims to isolate your AI traffic from Search Console, it was invented. The same scepticism applies to the SEO versus Google Ads decision, where the honest comparison is about time horizon and budget shape.
If you want the deeper version of the AI half, how to show up in ChatGPT and AI search is the full walkthrough, and my guide to llms.txt covers the one file everybody is currently being upsold. If the site itself is the bottleneck, the AI and systems assessment is $300 and credited in full against a build, and a custom build starts at $2,000 for a Trailhead and $6,500 for an Engine. Fill out the contact form with the query you wish you won, and I will tell you which fight you are actually in.
Sources and further reading
- Google Search Central: AI features and your website
Google documents that no additional requirements or special optimizations are needed for AI Overviews and AI Mode, because the generative features run on the core Search ranking systems.
- Google Business Profile: improve your local ranking on Google
Relevance, distance, and popularity, plus the direct statement that the categories you select affect your local ranking. This is the page that governs the map pack half of the fight.
- Pew Research Center: users are less likely to click links when an AI summary appears
Browser-tracked study of 900 US adults across 68,879 searches. Link clicks fell from 15 percent to 8 percent when an AI summary was present. US only, with no local-intent breakout.
- SEO.com: ChatGPT search statistics
Source of the overlap figure quoted here. Published by a company that sells AI search services, which is worth holding in mind while you read it.
- Goodfellas Tech: the AI search survival guide for small businesses
Source of the citation figure quoted here, from the same vendor-published family of analyses. Useful for direction, thin as proof.
Frequently asked questions
Are Google ranking and AI visibility the same thing?
No. They share a foundation and reward different finishing work. Google ranks whole pages against each other for a position in a list of links. An AI assistant assembles one answer from several sources and credits a few of them, so it rewards the clearest quotable passage rather than the strongest page overall. A business can do well at one and poorly at the other.
How much do Google results and AI citations overlap?
Less than most people assume, though the public numbers all come from vendors. One 2026 analysis published by SEO.com found the overlap between Google top ten results and what AI engines cite has fallen below 20 percent, and a related write-up from Goodfellas Tech reports roughly 90 percent of ChatGPT citations come from pages outside Google top 20. Both are published by companies selling AI search services. Treat them as direction.
If I win at SEO, do I automatically win at AI answers?
Not automatically. Ranking well means your pages are getting retrieved, which is the hard half. Whether one gets quoted depends on whether the answer sits in a clean, self-contained passage a model can lift. Plenty of well-ranked pages lose the citation to a worse-ranked page that simply answered the question in its first paragraph.
Does Google need special work to appear in AI Overviews?
No. Google states there are no additional requirements and no special optimizations necessary to appear in AI Overviews or AI Mode, because those features run on the core Search ranking systems. That makes AI Overviews the one AI surface where ordinary search fundamentals are the documented strategy.
Are AI answers taking away my clicks?
Some of them. The one credible study is from Pew, which tracked 900 US adults across 68,879 searches and found link clicks fell from 15 percent to 8 percent when an AI summary appeared. It is US only and has no local-intent breakout, so read it as a direction of travel. For a local business, a short search still lands in the map pack, where the call comes straight off your profile.
Which fight should a small local business prioritise?
Start with the shared foundation, because it is the only work that counts twice: a fast site with real HTML, one clear job per page, consistent details everywhere, and real reviews. Then take the map pack seriously, since short local searches still resolve there. Then make your pages quotable, which costs an afternoon of rewriting and helps skimming humans just as much.
How do I measure AI visibility?
By hand. Search Console folds AI Overviews traffic into the regular Web search type, so there is no separate report, and anyone selling you an AI traffic breakdown from Search Console is inventing it. Ask your key questions in ChatGPT, Perplexity, and Google AI Mode once a month, record who gets named and cited, and watch the list change.
What does Kootenay Made Digital charge to sort this out?
If you already have a site, the AI and systems assessment is $300, credited in full against a build, and it separates the fixes that pay into both fights from the ones that do not. If you do not have a site yet, custom builds start at $2,000 for a Trailhead and $6,500 for an Engine, and the foundation described in this guide is how I build by default.
Kootenay Made Digital
We build websites, local presence, and calm AI setups for Kootenay small businesses. No jargon, no agency fog, no surprise fees.
