Why Don't AI Tools Recommend Your Business Yet?
A prospective client typed a question into an AI tool last week, something close to “who is the best XYZ firm in Switzerland,” and your business never came up. Not because you don’t do the work. Not because your website ranks poorly on Google. It’s because the question never touched Google at all.
That’s the shift. Monthly visits across generative AI platforms grew 70% year-over-year, reaching 9.5 billion between June 2025 and May 2026 (Similarweb, 2026 Generative AI Landscape report). Every one of those visits is a research moment your Google ranking cannot reach, because the buyer never opens a search results page. They ask an AI assistant directly, and they act on whatever answer it gives them, including which businesses it recommends, and which ones it never mentions.
For most businesses, the honest answer to “does my business come up in that conversation” is: rarely, inconsistently, and in a way no one on the team can currently see. And if your business is one of the ones being recommended today, that position isn’t permanent either. AI platforms re-evaluate their sources constantly, so staying visible takes the same continuous work as winning visibility in the first place, small, compounding reinforcements every week rather than one project you finish and file away. Take ChatGPT, the largest of these platforms, as one measurable example: its rate of citing outside sources in its answers rose from about 1.6% in June 2025 to roughly 6.8% by May 2026, more than four times in eleven months. That growth is happening across AI tools generally, not on one platform alone, but the citation rate is not even across industries. Travel and hospitality businesses get cited in roughly 23% of relevant answers, automotive close to 20%. Professional services, the category most consulting, marketing, and B2B firms sit in, stays under 4%.
That gap is not a reason to wait. It’s the opening. A channel growing this fast, with this little competition inside your own category, rewards whoever gets structured for it first, and quietly drops whoever assumes their Google ranking still covers it.
What Changed: Why GEO Matters Now
Generative Engine Optimization, GEO, is the practice of structuring a website so AI tools like ChatGPT, Gemini, and Perplexity can understand, extract, and recommend the business, the same way SEO structures a site for Google. The two disciplines answer different questions. SEO earns a click. GEO earns a mention inside an answer the buyer never scrolls past, often the only answer they read before deciding who to call.
A website can rank first on Google and still be invisible in that answer, because ranking and being cited are two different tests. Google also reads a page and tries to understand what it means, then turns that understanding into a ranked list of many results, so a page it only partly understands can still show up, just further down. An AI model uses that same kind of understanding to write one synthesized answer, and names a business in it only when the page is specific and clear enough to be trusted as fact, not just relevant enough to rank. There’s no page two: a page that leaves the model unsure isn’t ranked lower, it’s left out of the answer entirely.
Most business websites were never built to be read that way. They were built to look right to a visiting human and to rank for a handful of keywords. The copy is written in language a person decodes using context clues, the logo, the industry, the tone. An AI model has none of that context. Working from the page alone, it either has to guess at what the business does, who it’s for, and why someone should choose it over the next result, or it skips the business and cites a competitor whose site already answers those three questions in plain language.
Why Your Website Needs to Be Alive And Never Finished
Most business websites are treated as a project with an end date: build it, launch it, revisit it once a year if at all. That worked when Google indexed a page once and mostly left it alone. It doesn’t work for GEO, because generative AI platforms re-crawl and regenerate their answers continuously, and a page that hasn’t changed in eight months carries none of the freshness signal that tells a model the information is current. A competitor who publishes a new FAQ answer this month, or updates a service page this week, becomes the more current, and therefore more trustworthy, source the next time that question is asked. A site that was “finished” in March is being quietly replaced in the answers by May.
This is also why a single optimization pass doesn’t hold. AI models change how they weigh sources, buyers change how they phrase questions, and competitors publish new content into the same gap that was just closed. A website that stays live, reviewed and updated on a fixed schedule instead of left alone after launch, keeps answering the questions buyers are asking this month, not the ones they were asking when the site went live.
What AI Actually Scans When It Answers a Question About Your Business
When an AI model is asked a question that touches a business, it is working from a specific set of signals, not a general impression of the brand:
Structured data and schema markup: Organization, Service, and FAQ schema translate a page into a machine-readable answer to “what is this business, what does it offer, and to whom.” A page without it forces the model to guess.
Direct answers to three questions: what do you do, who is it for, and why you and not someone else. If a page can’t answer all three in language a stranger could quote back, the model treats the business as unclear and moves to a source that isn’t.
FAQ content that mirrors real buyer questions: not marketing copy reformatted as a question, but the phrasing a buyer actually types into an AI tool.
Freshness: when the page was last meaningfully updated, and whether that update shows up on the page itself, not only in a file’s metadata.
Consistency across the wider web: whether directories, reviews, partner sites, and press mentions describe the business the same way the website does. A model that finds three different descriptions of the same company trusts none of them fully.
How to Make Sure Your Business Shows Up
Each scanning factor above has a direct fix, and none of them require rebuilding the site from scratch:
Add Organization, Service, and FAQ schema to every page that describes what the business does, not only the homepage.
Rewrite the copy on the core pages so it answers the three core questions, what you do, who it's for, and why you, within the first two sentences, in plain language, with no jargon standing in for a claim.
Build an FAQ section from the actual questions clients ask on discovery calls, in their words, not a rewritten service list.
Put a visible update on the site at least monthly, such as a new FAQ answer, a new case example, or an updated figure, so both the page and its metadata show recent activity.
Check what Claude, ChatGPT, Gemini, and Perplexity currently say when asked about the business’s category, and close the gap between that answer and what the site actually says.
How We Do This at Alldis & Partner
We run this as monthly steps, not a one-time project:
One GEO Improvement Sprint each month, one change to how the site is structured for AI, the one that month’s retest shows matters most, not a generic checklist.
Page and Content Updates, the messaging and clarity work that keeps the site answering the three core questions the way buyers are asking them this month, not the way they were asking eight months ago.
Retest and Recap, we ask the AI tools your buyers are already using what they say about the business, and report back exactly what changed, in plain language, every month.
That loop is the difference between a website optimized once and a website that stays visible as the AI models, the questions, and competitors’ content keep moving.
What It Costs You
This work runs at CHF 600 per month. Set that beside the cost of doing nothing: every month a site isn’t structured for AI, it stays absent from a growing share of the roughly 9.5 billion monthly visits happening across generative AI platforms, and every month a competitor updates their FAQ page first, they take the mention that would otherwise have gone to you. The cost of staying invisible in a channel your buyers are already using compounds quietly, off any analytics dashboard, until a client mentions, months later, that they “just asked AI.”
Is This You?
This is built for businesses that:
Want to know exactly what Claude, ChatGPT, Gemini, and Perplexity currently say about them, whatever their current Google ranking looks like
Want to be the business AI recommends when someone asks about their category
Don't have the internal time or discipline to review and update the site's AI-facing content every month
See What AI Currently Says About Your Business
You can keep assuming the Google ranking covers this. Or find out this month, exactly, what Claude, ChatGPT, Gemini, and Perplexity tell someone asking about your category, and fix the specific gaps keeping the business out of that answer.
Sources
Similarweb, “2026 Generative AI Landscape” report: generative AI platform visit growth (70% YoY; 9.5B visits, June 2025–May 2026), ChatGPT citation rate growth (1.6% → 6.8%), and citation rate by industry (Travel & Hospitality ~23%, Automotive ~20%, Professional Services <4%): similarweb.com/blog/marketing/geo/gen-ai-stats
Alldis & Partner GEO service page (definitions, three-question framework, service structure, pricing): alldis.marketing/en/geo
