The Recommendation Gap: why AI doesn't name your business, even when Google does
You can rank on the first page of Google and still never be named by ChatGPT. The two are different systems, answering different questions, and the gap between them is where a growing share of your customers now make their decision.
We call it the Recommendation Gap: the distance between being findable and being recommended. Findable means a search engine will show your page in a list if someone looks for it. Recommended means an AI assistant names your business, unprompted, when a buyer asks who to use. Most businesses have spent fifteen years closing the first gap. Almost none have noticed the second one opening.
The symptom you cannot see
The reason the Recommendation Gap goes unnoticed is that it is silent. When a buyer types "best [your service] near me" into ChatGPT, Claude, Gemini or Perplexity, they are handed a short list of specific businesses. If you are not on that list, they contact one of the names they were given. You were never in the conversation, and nothing about it reaches you. No missed call, no bounced visit, no line in your analytics.
We call this Silent Shortlisting. The buyer was shortlisting suppliers in your category, in your town, at that moment, and you were quietly left off. Traditional analytics cannot show you an enquiry that was decided before anyone reached your site. The first time most owners see it is when they run a scan and watch three competitors get named and their own business get a blank.
Findable is not recommended. A page that ranks can still be invisible to the system that increasingly makes the choice.
Why the gap exists: the Three Gates
An AI assistant does not rank pages the way a search engine does. Before it will name a business, it has to pass that business through three gates. Fail any one of them and you are not recommended, no matter how well you rank.
Gate one: Retrieval. Can the engine find and read you at all? This is the technical layer. Is your content served as real text rather than only rendered by JavaScript. Do you allow the AI crawlers. Is there structured data, an llms.txt file, clean headings. If the engine cannot retrieve and parse you, nothing else matters. This is the gate most "AI SEO" advice stops at, and it is necessary, but on its own it is rarely the reason a business is missing.
Gate two: Resolution. Can the engine work out who you are, precisely, and trust that it has the right entity? This is entity clarity. Does your site state plainly what you do, where, and for whom. Is your business one consistent, well-described entity across your own pages and the profiles you link to. An engine that cannot resolve you into a confident, single answer will leave you out rather than risk naming the wrong business.
Gate three: Corroboration. Does the wider internet agree that you are a real, credible answer? This is third-party authority: reviews on the platforms engines trust, mentions in "best of" articles, forum threads, directory listings, press. An assistant recommending a supplier is staking its answer on you. It wants the rest of the web to back it up before it does. This is the gate most small businesses actually fail, and it is the one no amount of on-site work can close on its own.
Where the effort usually goes wrong
Almost all of the money spent on "getting into AI search" goes into gate one. Schema markup, an llms.txt file, FAQ pages, tidy metadata. That work is real and worth doing. But retrieval is table stakes, not a differentiator. A business can have a technically excellent site and still be absent from every answer, because it never cleared corroboration.
The reverse is also true, and more instructive. In competitive categories we have seen businesses with almost no technical readiness get named as the top pick, because the wider web triangulates them clearly: strong reviews, niche forum mentions, specialist coverage. The engine did not need schema to find them. It needed the third-party signal, and they had it.
The lesson is consistent. Technical readiness amplifies authority. It does not create it. If your category is competitive and your corroboration is weak, more schema will not save you. If you are spending on gate one while gates two and three sit open, you are polishing the wrong lever.
How to close the Recommendation Gap
Closing the gap means working all three gates, in order of what is actually missing rather than what is easiest to buy.
First, measure it. Ask the engines the questions your customers ask and see who gets named instead of you. That is what an AI Referral Score is for: a plain reading of how often ChatGPT, Claude, Gemini and Perplexity recommend you, scored against a starting baseline so you can tell whether anything you do moves it.
Second, clear retrieval and resolution. Make the site readable, structured, and unambiguous about who you are. This is the fast, controllable part, and it is the foundation everything else sits on.
Third, and this is where the gap usually lives, build corroboration. Grow real reviews on the platforms the engines read. Earn genuine mentions in the places your category is discussed. Publish substance that other people have reason to cite. This is slower, it happens off your own site, and it is the work that actually changes who gets recommended.
The engines re-weight on their own timeline, typically over sixty to ninety days, which is why the Recommendation Gap rewards patience and punishes one-off checks. A single scan tells you almost nothing. A baseline you measure against, month after month, tells you whether the gap is closing.
Findable was the last fifteen years. Recommended is the next fifteen. The businesses that notice the gap first, and work the right gate, are the ones AI will be naming while their competitors are still wondering why the phone went quiet.