We ran this test on ourselves — here are the ugly numbers
In August 2026 we asked ChatGPT the same 25 questions a potential customer would ask — about our own company. The results: zero organic recommendations out of 25. Asked "who does AI search optimization in Southern California?" — our exact business, our exact market — it listed nine competitors and not us. Worse: when we asked about our brand by name, three of the four answers described the wrong company entirely — attributing our business to an unrelated firm in North Carolina with a similar name, and once to a fitness lead-gen outfit.
We're publishing that so you'll believe the next sentence: this is normal. Almost every small business is invisible in AI answers right now — not because AI hates small business, but because the signals engines need are missing. That's bad news with a silver lining: the fixes are known, they're mechanical, and almost none of your competitors have done them either.
Reason 1: there's nothing on your site an AI can quote
AI engines compose answers from passages they can lift whole. If your homepage opens with "Welcome to our website" and the actual answer to "what do you do and for whom" lives in paragraph nine — or nowhere — the engine quotes a competitor whose page answers in the first screen. The fix is structural, not poetic: every important page should answer its one question directly, near the top, in a self-contained block a machine could quote without context.
Reason 2: your facts disagree across the internet
Engines cross-check before they recommend. When your Google listing, Yelp page, Bing listing and website disagree about your hours, your category, or even your name's spelling, the engine doesn't pick a winner — it hedges, or skips you. An afternoon of listing cleanup (same name, same category, same hours, everywhere) is some of the highest-leverage AI-visibility work that exists, and it costs nothing but attention.
Reason 3: you share a name with someone — and the AI can't tell you apart
This is the one that bit us. If any company anywhere shares your name (or nearly), AI engines can blend you together — answering questions about you with their location, their reviews, their reputation. The cure is entity disambiguation: machine-readable identity records (structured data on your site, a knowledge-base entry, consistent registered identifiers) that let an engine prove which business is which. It's invisible work — nobody sees a schema block — but it's the difference between "here's that company" and "here's some company with that name."
Reason 4: your site literally blocks AI
A surprising share of small-business sites block GPTBot, PerplexityBot and their peers in robots.txt — sometimes from an old "block all bots" snippet, sometimes because a plugin did it silently. If the crawler can't read your site, nothing else on this list matters. Checking takes one minute: look at yourdomain.com/robots.txt for those bot names under a Disallow rule.
Reason 5: no third-party corroboration
Engines weight what others say about you at least as heavily as what you say about yourself — reviews on platforms they trust, mentions in directories and articles, registries that confirm you're real. A business with zero footprint beyond its own site asks the engine to take its word for everything. Engines don't.
What fixing this actually looks like
The order matters: identity first (listings agreement + entity records + the name-collision check), quotable answers second (rewrite money pages answer-first), corroboration third (reviews, registries, mentions), and measurement throughout — the same fixed panel of questions, re-asked monthly, so you know whether any of it worked. That last part is the piece almost everyone skips, which is why almost everyone's "AI SEO" is a feeling instead of a number. We run the exact playbook above on our own business — same baseline, same fixes, same monthly panel — and publish what we find.