Why GEO matters now: the case without the borrowed statistics

One: an answer has room for few names

A results page lists every business that matches. An assistant's answer to "who should I call" names a few and stops, so each place in it is scarce, and there is no second page to fall back to. In the probe that accompanied our audit of 9,483 small business websites, one buyer question per business went to Claude with web retrieval on, and 14.9% of 9,240 businesses were named; among the answers that gave a clearly differentiated list, 17.7%. Most businesses were outside the answer for a plain question about their own trade in their own town.

That is one question, one assistant and one day, so read it as a direction. It is still the fact that makes the rest matter: when the list is short, being on it or off it is the whole result.

Two: most sites leave the facts out

The same audit counted what a crawler received from those sites. 61.5% had no question-and-answer content of any kind. 43.1% gave more than one account of who and where they are across their own homepage, contact page and about page. 64.7% showed a crawler no review signal, and 17.8% had no schema.org markup at all. 7.2% turned the crawler away with a challenge page.

Read the other way, those figures are the opportunity. The field is thin: in most trades, most competitors have not yet stated their services, prices, hours and area in a form an answer can use. A business that does it now is doing work few others in its market have done, and the answer can only draw on what someone wrote down.

Three: the work is mostly work you should do anyway

Letting crawlers in, stating one name, address and phone everywhere, marking up the business, putting the facts in the page's text and answering buyers' questions in the first sentence under each heading: every one of those helps search as well as answers. Google describes AI Overviews and AI Mode as rooted in its core Search ranking in its guide to generative AI search, and its page on AI features says they need no special markup or file. The part that is only for AI answers is small, and GEO vs SEO, task by task sorts it out.

So the downside is modest. If the assistants turned out to matter less to your buyers than you expected, the work would still have made the site clearer, more consistent and easier to find in search.

Four: you can find out where you stand in an afternoon

No part of this has to be taken on trust. Write six of your buyers' questions the way they would type them, put each to ChatGPT, Claude, Gemini and Perplexity three times, and write down who is named. Run the crawler access check on your site. By the end of the afternoon you know whether you are in the answers, who is in them instead, and whether the crawlers can reach you. Setting up prompt monitoring turns that into a monthly reading.

What nobody can tell you

How many of your own buyers ask an assistant before they call. Published usage figures describe the assistants' whole audiences, not your trade in your town, and they change fast. The honest way to know is to ask new customers how they found you, with an AI assistant among the answers, and read the replies over a few months. Calculating the ROI of GEO shows how to turn that count into a break-even test for the work.

Nor can anyone promise that the work puts you in the answers. The choice is made inside the assistants. What the work does is remove the reasons an answer would leave you out, and the monthly reading shows whether it did.