How Does ChatGPT Decide Which Business to Recommend? (It's Not What You Think)
One of the first questions I get after delivering an AI visibility audit is: "How does ChatGPT decide who to recommend?"
It's a fair question. Most business owners assume it works like Google — whoever has the best SEO wins. But that's not how AI platforms operate. And the difference explains why some businesses with perfect Google rankings have 0% visibility on ChatGPT, while competitors with half the reviews show up every time.
Here's what I've learned from running audits across ChatGPT, Gemini, Perplexity, Claude, and Google AI Overview.
AI doesn't rank. It recommends.
Google gives you ten blue links and lets you choose. AI is different — it picks two or three businesses and presents them as recommendations. That's a fundamentally different process. Google is a librarian pointing you to a shelf. AI is a trusted friend saying "go with these guys."
Because of that, the signals AI uses are different from the ones Google uses.
Signal 1: Can AI actually read your website?
This is the biggest gap I find in audits. A business can have a beautiful website that ranks well on Google, but if it's missing structured data — specifically schema markup — AI platforms struggle to parse it.
Schema markup is invisible code that tells AI what your business is, what services you offer, where you're located, and how to contact you. Without it, AI has to guess. And when AI guesses, it often guesses wrong or just skips you entirely.
I recently audited a childcare centre with great reviews and a solid local reputation. Their robots.txt file was actively blocking ChatGPT's crawler from reading the site. The business was literally telling AI to go away — and their visibility score reflected it.
Signal 2: What do your reviews actually say?
AI doesn't just count stars. It reads review text. Platforms like ChatGPT and Perplexity scan what customers are actually writing — specific services mentioned, location references, quality descriptors.
A business with 30 reviews that consistently mention "fast response time" and "great with commercial clients" will often outrank a competitor with 200 generic five-star reviews. AI values specificity because it needs to match your business to specific customer queries.
This is why review automation matters — not just for volume, but for guiding customers to mention the services and qualities that AI is looking for.
Signal 3: Are you mentioned outside your own website?
AI cross-references sources. If your business name, address, and phone number appear consistently across directories, review platforms, and industry sites, AI treats you as more trustworthy. If the information is inconsistent — different phone numbers on different directories, or a different business name on Yelp versus Google — AI gets confused and may exclude you from recommendations entirely.
This is the same concept as traditional citations in SEO, but AI weighs it differently. It's less about the number of citations and more about consistency across them.
Signal 4: Do you have content AI can cite?
When AI recommends a business, it often explains why. "This firm specialises in X" or "they're known for Y." That explanation has to come from somewhere.
If your website has a FAQ section that answers common customer questions in natural language, AI can pull directly from it. If you have blog posts that demonstrate expertise in your field, AI treats that as authority. If your service descriptions are detailed and specific rather than generic, AI has something meaningful to reference.
Generic content like "We are a leading provider of quality services" gives AI nothing to work with. Specific content like "We specialise in FBT compliance for salary packaging, with dedicated services for car parking valuations across Sydney, Melbourne, Brisbane, and Perth" gives AI exactly what it needs.
Signal 5: Does Google's AI Overview show you?
This one is newer but increasingly important. Google's AI Overview — the AI-generated summary that now appears above search results for many queries — pulls from Google's own index. If you're not appearing in AI Overviews for your core service queries, that's a signal that your content isn't structured for AI consumption, regardless of your traditional search ranking.
I test every audit against Google AI Overview alongside the other four platforms. It often tells a different story than ChatGPT or Perplexity, which means optimising for one platform isn't enough.
The bottom line: AI visibility is a different game from SEO.
You can rank #1 on Google and be completely invisible to ChatGPT. I've seen it multiple times across industries and countries. The businesses winning on AI platforms aren't necessarily the ones with the biggest marketing budgets — they're the ones whose online presence is structured in a way AI can actually read, understand, and cite.
The good news is that most of these signals are fixable. Schema markup, FAQ content, review strategy, directory consistency — none of this is rocket science. But it does require knowing what to look for, and most business owners don't know it's a problem until someone shows them.
If you're curious where your business stands, I run free AI visibility audits across all five platforms. Takes about 15 minutes on my end, and you'll have the results within 48 hours.
— Chris Small Founder, Smartech Marketing Systems
