Why You Rank on Google But Don't Exist in ChatGPT

Take twenty seconds and run this test. Open ChatGPT. Type: best [what you do] in [your city]. Read the answer.

Are you in it?

Most business owners we ask have never tried this. They check their Google ranking, see themselves on page one, and reasonably assume they are covered. Then they run the test and find three competitors named — none of them themselves.

That gap is not a ranking problem. It is a different problem, and the numbers say it is getting expensive fast.

The channel nobody was watching just became the best one

Adobe Analytics tracks over a trillion visits to US retail sites. Their Q1 2026 report found that traffic arriving from AI assistants grew 393% year over year, peaking at 1,151% growth in December during the holiday season. Travel sites saw 233% growth, financial services 158%.

Volume alone would be easy to dismiss as curiosity. The conversion data is not.

In March 2025, visitors arriving from AI assistants converted 38% worse than everyone else. Twelve months later, in March 2026, they converted 42% better — a complete reversal. Same channel, same stores, different year. Those visitors also spent 48% more time on site, browsed 13% more pages, and produced 37% more revenue per visit than traffic from any other source.

Read that sequence again, because it is the whole argument. The channel went from worst-performing to best-performing in a year, while growing at triple digits.

One honest caveat: this is US retail data, measured by Adobe on its own analytics platform, and it is not independently audited. Shopify's separately reported commerce data points the same direction. Treat the exact figures as indicative and the trend as real.

Two machines, two jobs

A search engine indexes pages and returns a list. Ten blue links, a map pack, some ads. The person searching does the work of choosing.

An answer engine does the choosing for them. Ask ChatGPT, Perplexity, or Google's own AI Overviews for a recommendation and you do not get ten options. You get two or three names, in a sentence, with reasons attached. The list has already been filtered.

Here is the part that matters: the shortlist is not the top of the ranking. These systems do not crawl a keyword and read off positions. They assemble an answer from what they understand about businesses — what is written about you, where you are mentioned, whether your own pages state plainly what you do, how much third-party evidence exists that you are real and competent.

So you can hold position three on Google and be absent from the answer entirely. Not ranked lower. Absent.

Why you will never see this in your analytics

This is what makes the problem genuinely dangerous rather than merely annoying.

When you lose a Google click, you see it. Impressions stay flat, clicks drop, position slips — the loss shows up in Search Console as a number that got smaller.

When you lose an AI recommendation, nothing happens. There was no impression. There was no click to miss. Someone asked for the best option in your category, heard three other names, and went to one of them. Your dashboards record absolutely nothing, because from your side of the glass the event never occurred.

Every other marketing problem announces itself. This one does not. Which is why most businesses discover it by accident, usually from a customer who mentions they asked an AI assistant and got sent somewhere else.

A third of the average homepage is invisible

Adobe also built a tool that scores how much of a web page a language model can actually read. Their retail benchmark found that 34% of homepage content was invisible to AI models. Product pages averaged 66% readable. Meaning: on a typical retail site, a third of what a visitor sees simply is not there when a machine looks.

The pattern in which pages scored well is the useful part. Returns pages, Contact Us pages and FAQ pages all scored above 80% — the plainest, most text-heavy pages on the site. Homepages and category pages, the designed ones, scored lower.

That inverts how most businesses allocate effort. The pages you spend least time on are the ones machines read best, and the showpiece homepage is where content goes missing.

What actually makes a business legible to an AI engine

Six things carry most of the weight. None of them are tricks.

• Your pages have to say what you do, in plain words, in the source. This catches modern sites hardest. Many websites built in the last two years render entirely in the browser — beautiful in Chrome, and nearly empty when a crawler or a model fetches the page. If the HTML that arrives contains a logo and a loading state, that is what the machine learns about your business. We have audited sites where the visible page had 900 words and the source had eleven.
• Structured data. Schema markup is how a page tells a machine what kind of thing it is — a local business, a service, a set of questions and answers, a price. Without it, models infer from context and often infer wrong. It costs nothing to add and almost nobody in Indian SMB web work does it.
• Mentions on sites that are not yours. This is the biggest shift from classic SEO. For years the currency was backlinks. For AI visibility, being talked about matters more than being linked to — directory listings, local press, industry roundups, forum answers, supplier pages. A model that has seen your name in twenty places knows you exist. One that has only seen your own website has a single source, and a single source is thin.
• Reviews, in volume and recently. Review count is one of the few hard numbers these systems can lean on to justify a recommendation. Forty recent reviews beats four excellent ones from 2023. Unglamorous, and it works.
• A consistent business identity everywhere. Same name, address, phone and category across Google Business Profile, Justdial, your site, your social bios. Inconsistency does not just confuse customers — it makes a model uncertain which entity you are, and uncertainty loses shortlists.
• Content that answers questions the way people ask them. Not keyword pages. Actual questions, answered directly, near the top. What does this cost? How long does it take? What is the difference between X and Y? This is exactly why FAQ pages score above 80% on readability while homepages do not.

Three things worth doing this week, free

You do not need an agency for any of this.

• Run the test properly. Try five phrasings, not one — best, affordable, near me, recommend, and a problem-shaped query like who can fix X in Bengaluru. Note which competitors keep appearing. That list is your real competitive set, and it may not match the one you have been watching.
• View your own page source.Right-click, View Page Source, and read what is actually there. If you cannot find your service names and your city in that text, neither can a machine. Given Adobe's 34% figure, assume you have a problem until you have checked.
• Claim and complete every listing you can find. Google Business Profile first, fully filled, with categories and photos and hours. Then whatever directories exist for your sector. Boring work, and the cheapest visibility you will ever buy.

Where this is heading

The direction is not subtle. Buyers who used to open a search results page now open a chat window, and Adobe's own consumer survey found 39% have already used AI for shopping — with 85% of those saying it improved the experience. Retail moved first because retail always does. Travel and financial services followed within quarters. Local services are next.

None of which means Google stops mattering. It means the thing you have been optimising is now one of two doors, and most businesses have only ever checked the first.

The ones who get ahead here will not be the ones with the biggest budgets. They will be the ones who ran the test early, found out they were invisible, and fixed the six unglamorous things above while their competitors were still checking keyword positions.

Data in this article: Adobe Analytics, Q2 2026 AI Traffic Report (published 16 April 2026, covering calendar Q1 2026) and the May 2026 update, based on more than one trillion visits to US retail sites, plus Adobe's consumer survey of over 5,000 US consumers.