Series: Getting Cited by AI — Post #2. Read Post #1: How to Get Your Business Cited by ChatGPT →
You typed your own service and city into ChatGPT — "best wedding photographer in Nashville," "roof repair near Austin," "med spa in Phoenix" — and the AI confidently named three businesses. None of them were you. One of them was the competitor you know you out-deliver.
That stings, but it isn't a popularity contest you lost. It's a retrieval problem you can fix. AI models don't rank the "best" business — they recommend the one whose information is easiest to read, easiest to trust, and easiest to quote. Here's why your competitor keeps winning that, and what to change.
AI isn't picking the better business. It's picking the more legible one.
When you ask an AI assistant for a local recommendation, it isn't forming an opinion. Under the hood it's doing some mix of: pulling from what it learned in training, retrieving live pages, and pulling from structured sources (maps, directories, review platforms). Then it stitches a confident-sounding answer from whatever it can cleanly extract.
So the business that gets named is rarely "the best." It's the one that is:
- Findable — its pages get crawled and retrieved.
- Parseable — its key facts (what it does, where, hours, price range, specialties) are in plain text and structured data, not buried in an image or a slider.
- Corroborated — the same facts show up consistently across its site, Google Business Profile, and directories.
- Reviewed — it has enough recent, specific reviews that the model treats it as a safe recommendation.
Your competitor probably didn't out-market you. They just happen to clear those four bars — often by accident, because they used a template that ships with schema, or because they have more Google reviews. Let's go through why each one tips the scale.
Reason 1: Their facts are in text. Yours are in pictures.
The single most common reason a great business is invisible to AI: its most important information lives inside images, video, or design elements that models can't read.
Your services are a graphic. Your pricing is a PDF. Your specialties are baked into a hero banner. Your "About" is a beautiful photo with three words on it. A human gets all of this instantly. A language model gets… almost nothing.
Your competitor's site might be uglier — but if it says, in actual selectable text, "We're a family-owned roofing company in Austin, Texas, specializing in storm-damage repair and metal roofs, serving Travis and Williamson counties since 2009," then the AI has a clean, quotable sentence. Yours has a JPG.
Fix: Audit every page for facts that exist only as an image. Put them in real text too. You can keep the pretty version — just make sure the same information also exists as words the page can render. Aim for one plain sentence that answers "what, where, and for whom."
Reason 2: They have schema markup. You don't.
Schema (structured data) is a hidden block of code that labels your facts for machines: this is the business name, this is the price range, these are the hours, this is a review, this is an FAQ answer. It's the difference between handing a model a labeled spreadsheet versus a wall of prose.
Many website builders and templates now include basic schema automatically — which is exactly how a less-polished competitor ends up "more legible" than your custom-built site that has none. They didn't do anything clever. Their template did it for them.
Fix: Add LocalBusiness schema with your name, address, phone, hours, and price range. Add FAQPage schema to your FAQ. If you collect reviews on your own site, mark them up too. On Wix/Squarespace/WordPress you can inject this with a code block or a schema app — no rebuild needed. (Post #3 in this series will give copy-paste templates.)
Reason 3: Their story is consistent. Yours contradicts itself.
AI models hedge against businesses whose facts don't line up. If your website says you're in one suburb, Google says another, an old directory lists a disconnected phone number, and your Instagram bio says something different again — the model sees noise, lowers its confidence, and recommends the business it can describe without contradicting itself.
This is the quietest killer because nothing looks broken. Every individual page is fine. It's the inconsistency across them that demotes you.
Fix: Pick your canonical facts — exact business name, address, phone, primary service — and make them identical everywhere: website footer, Google Business Profile, Yelp, industry directories, social bios. Same spelling, same format. Kill or correct stale listings. Consistency reads as trust.
Reason 4: They have 40 recent reviews. You have 6 from 2022.
Reviews are one of the strongest signals feeding AI local answers, because they're independent corroboration the model can lean on. A business with a steady stream of recent, specific reviews ("they fixed our leak the same day," "great with nervous brides") gives the model concrete, quotable detail. A business with a handful of old, generic reviews gives it nothing to say — so it stays quiet about you and talks about your competitor instead.
It's not just the star rating. It's volume, recency, and specificity.
Fix: Build a simple, repeatable ask — a text or email to every happy customer right after you deliver, with a direct link to your Google review form. Nudge them to mention the specific thing you did. Twenty fresh, detailed reviews over a couple of months will move you more than almost any on-site tweak.
Reason 5: They answer the question. You describe yourself.
Most small-business websites are written like brochures: "Welcome. We're passionate about excellence. Our team brings years of experience." That's invisible to AI because nobody asks an assistant "who is passionate about excellence." People ask specific questions: "Who does same-day roof repair in Austin?" "Which Nashville photographer shoots elopements?" "Is there a med spa in Phoenix open on Sundays?"
The business that gets recommended is the one whose pages literally contain the answer to those questions in plain language. Your competitor's FAQ probably says "Yes, we offer same-day emergency repair across Austin." Yours says "Quality you can trust."
Fix: Write the actual questions your customers ask into your pages — as headings and FAQ entries — and answer them in one or two direct sentences each. Answer-shaped content is quote-shaped content.
The 15-minute reality check
Before you change anything, see what AI says about you right now:
- Open ChatGPT (and one other assistant) and ask the exact question a customer would: "best [your service] in [your city]."
- Note who gets named — those are your real AI competitors, regardless of who you think they are.
- Ask directly: "What do you know about [your business name]?" Read what comes back. Vague, wrong, or "I don't have information" all mean the same thing: you're not legible.
- Open a named competitor's website. View its source and search for
application/ld+json— that's their schema. Check whether their facts are in text and whether they have more recent reviews.
That comparison usually makes the gap obvious. It's almost never that they're better. It's that they're readable and you're not.
The order to fix it in
You don't need all of this at once. Highest impact first:
- Reviews — start the ask today; it compounds slowest, so begin now.
- Facts in text — move image-only info into words this week.
-
Schema — add
LocalBusiness+FAQPage; a one-time job. - Consistency — align name/address/phone everywhere.
- Answer-shaped content — rewrite key pages as questions-and-answers.
Do the first three and most small businesses go from invisible to quotable.
Free offer: I'll run this exact comparison on your real site versus whoever AI currently recommends in your area, and send you a one-page snapshot — what AI says about you now, who it names instead, and the three highest-impact gaps to close — for free. If it's useful and you want the fixes done, we can talk. If not, keep the snapshot. No catch.
📧 Send your website URL to faithpath25@gmail.com with subject "GEO snapshot" and I'll send yours back.
FAQ
My competitor is objectively worse — how are they ranked above me by AI?
Because AI recommends the more legible business, not the better one. Cleaner structured data, more recent reviews, and facts written in plain text usually explain the whole gap.
Will fixing my site guarantee AI starts recommending me?
No — there's no guarantee and no fixed timeline. These changes make you readable and trustworthy to the models; pickup depends on when your pages are next crawled and how often the underlying sources refresh. They're the necessary groundwork, not a switch.
Is paying for ads the faster route?
Ads buy temporary visibility on ad networks; they don't make you quotable in an AI's organic answer. The legibility fixes here are what get you named when someone asks an assistant directly — and they keep working after you stop paying.
How many reviews is "enough"?
There's no magic number, but recency and specificity matter as much as count. A steady trickle of detailed, recent reviews beats a large pile of old generic ones.
Two ways to act on this:
🔎 Free, no-strings: send your site URL to faithpath25@gmail.com with the subject "GEO snapshot" — I'll send back a 1-page read of exactly what AI assistants can and can't currently see on your site, plus the specific fixes. Free pilot, wherever you operate; if it's useful, a short review is all I ask.
🧰 Do it yourself: the copy-paste schema kits, checklists, and the full GEO audit live at SprintLanding → (includes a free starter). Prices in USD; Gumroad converts to your local currency at checkout.
Top comments (0)