You've put in the work. Your site ranks on page one of Google for your main keywords. Traffic comes in. The SEO boxes are ticked. But when someone asks ChatGPT to recommend a business like yours, your name doesn't appear. Not even close.
This is one of the most common frustrations we hear from business owners right now, and it makes complete sense why it's confusing. If Google can find you, why can't ChatGPT? The answer comes down to a fundamental difference in how these two systems work, and understanding that difference is the first step to doing something about it.
Google and AI search engines are looking for completely different things
Google is a retrieval system. It crawls your pages, indexes them, and serves them up when someone types a matching query. Ranking well on Google means you've got the right keywords, a reasonable number of backlinks, decent page speed, and solid on-page signals. Google doesn't need to deeply understand your business. It just needs to match your content to a search query.
AI search engines like ChatGPT, Perplexity, and Gemini work differently. They don't retrieve pages in real time in the same way. Instead, they draw on large language models (LLMs) that were trained on vast datasets of web content, along with real-time retrieval in some cases. When someone asks ChatGPT to recommend a supplier, a local service, or a specialist product, the model has to confidently understand what your business does, what it specialises in, who it serves, and why it's trustworthy enough to mention.
That requires a very different kind of signal than a keyword-optimised blog post.
The signals that make you visible to AI are not the same as traditional SEO signals
Traditional SEO rewards things like keyword density, anchor text, domain authority, and click-through rates. None of these tell an AI model what your business actually is.
AI models look for structured, unambiguous information. Specifically:
- Schema markup and JSON-LD that explicitly defines your business type, products, services, reviews, and FAQs in a machine-readable format
- Clear, consistent entity signals across your site and across the wider web (directories, press mentions, social profiles, third-party citations)
- Factual, specific content that directly answers the kinds of questions people ask AI tools
- Structured data that connects the dots between what you sell, who you sell it to, and what makes you different from competitors
If your site doesn't have these signals, an AI model simply can't build a confident picture of your business. And if it can't build a confident picture, it won't mention you, even if your homepage ranks number one on Google.
Why schema markup matters so much for AI visibility
Schema markup is a vocabulary of structured data that you embed in your site's code. It tells machines, not just humans, what your content means. A product page might look great to a human visitor, but without schema markup, an AI model has to guess what it's looking at.
With proper JSON-LD schema in place, you can explicitly communicate things like:
- Your business name, type, and location
- The specific products or services you offer
- Your aggregate review score and number of reviews
- Your FAQ answers, which AI models actively pull from
- Your founding date, areas served, and price range
This is the kind of clean, structured information that AI systems can consume with confidence. It's the difference between a model saying "I'm not sure who the best option is here" and saying "Based on what I know, this business specialises in exactly that."
If you're not sure what types of schema your site should have, this guide on schema types for different businesses is a good place to start.
The entity problem: AI doesn't know who you are
One of the deeper issues is something called entity recognition. AI models build knowledge about the world through entities: named things with defined attributes. Google has its own knowledge graph full of entities. AI models have something similar baked into their training data.
If your business appears in enough places, with consistent information, in structured formats that AI systems can parse, you become a recognised entity. That means the model can refer to you with confidence.
If your business only exists as a collection of keyword-optimised pages with no structured data and minimal third-party mentions, you're essentially anonymous to the model. You might rank on Google, but to an LLM, you're noise.
This is why e-commerce brands with thousands of products but no schema markup are invisible to AI, even when they're well-established businesses with years of Google rankings behind them. The signals that built those Google rankings just don't transfer.
Content structure matters too, not just technical markup
It's not only about adding schema to your site. The way your content is written and structured also affects how well AI systems can extract useful information from it.
AI models are much better at processing content that:
- Directly answers specific questions rather than talking around them
- Uses clear headings and logical structure
- Includes specific facts, numbers, and named details rather than vague marketing copy
- Defines what the business does in plain, unambiguous language early on each page
A homepage that says "we're passionate about delivering world-class solutions" tells an AI model almost nothing. A homepage that says "we supply certified organic skincare to independent retailers across the UK, with a minimum order of 6 units and free delivery on orders over £150" gives the model something to work with.
This connects to the broader concept of generative engine optimisation (GEO), which is specifically about making your content more extractable and usable by AI systems. Our post on GEO vs SEO breaks down the distinction clearly if you want to go deeper on this.
Your robots.txt file might be blocking AI crawlers
Here's one that catches a lot of people out. Some AI systems do crawl the web in real time, and they use their own crawlers to do it. If your robots.txt file is blocking those crawlers, your site simply won't be included in their real-time retrieval.
Common AI crawlers include GPTBot (OpenAI), ClaudeBot (Anthropic), and PerplexityBot. Many sites have broad disallow rules that inadvertently block these bots without the site owner realising.
Checking your robots.txt file takes about two minutes. Go to yourdomain.com/robots.txt and look for any rules that might be blocking the crawlers mentioned above. If you see a blanket Disallow: / for any of them, you're cutting yourself off from real-time AI indexing entirely.
What you can actually do about it
The good news is that this is a solvable problem. It requires a different set of actions than traditional SEO, but it's not mysterious once you know what you're dealing with.
Here's a practical starting point:
- Audit your current schema markup. Most sites have none, or have very basic, incomplete schema that doesn't communicate enough. You need to know what you're working with before you can fix it.
- Add or fix your JSON-LD schema. At minimum, your business needs Organisation or LocalBusiness schema, plus Product, Review, and FAQ schema where relevant.
- Check your robots.txt file for rules that block AI crawlers.
- Audit your entity footprint. Are you listed in relevant directories? Do your Google Business Profile, website, and third-party mentions all use consistent information?
- Rewrite key pages to be more factual, specific, and directly useful. Swap vague marketing language for concrete details about what you offer, who you serve, and what makes you different.
- Consider an llms.txt file. This is a newer convention that gives AI models a clear, structured summary of your business and content. Learn more about llms.txt and whether you need one here.
If you're not sure where you currently stand, FlinnSchema's free AI visibility audit will show you exactly what's missing and what's having the biggest impact on your AI search presence.
The gap between Google ranking and AI visibility is widening
This isn't a temporary quirk that will sort itself out. The way people find businesses is changing. A growing number of buyers, particularly in the 25 to 45 age bracket, now use AI tools as their first port of call when researching suppliers, products, and services. If you're not visible in those systems, you're not in the conversation at all.
Google rankings are still valuable. Nobody is saying to abandon SEO. But they are no longer sufficient on their own. Businesses that build their AI visibility now, while most of their competitors haven't even realised the problem exists, are going to hold a significant advantage over the next two to three years.
The businesses that wait until AI search is mainstream before taking action will find the gap much harder to close. The time to get this right is while it's still relatively straightforward to stand out.
Frequently Asked Questions
If I rank on Google, doesn't that mean AI can find me too?
Not automatically. Google ranking relies on keyword and link signals that don't translate directly into AI visibility. AI models need structured, machine-readable information to confidently identify and recommend your business. A high Google ranking does not guarantee that an AI model knows what your business does or has enough confidence to mention it in a response.
How long does it take to become visible in AI search results after making changes?
It varies. For sites that are already being crawled by AI systems, adding structured data and improving content clarity can start to show results within a few weeks. For sites with very low entity recognition or that have been blocked by robots.txt rules, it can take longer as those systems need to re-crawl and re-process your content. Generally, expect to see meaningful improvement within one to three months of implementing the right changes properly.
Is AI visibility something I can sort out myself, or do I need a specialist?
Some of the basics, like checking your robots.txt file or improving the factual clarity of your homepage copy, you can absolutely do yourself. Implementing comprehensive, technically correct JSON-LD schema across a full product catalogue or a complex site is more involved and is where mistakes are easy to make. If you want to get it right efficiently, working with a specialist like FlinnSchema will save you a lot of trial and error.
Does having lots of online reviews help with AI visibility?
Yes, but only if they're structured correctly. Reviews that are marked up with proper schema (including aggregate rating and review count) give AI models a strong trust signal. Reviews that exist on third-party platforms but aren't reflected in your site's structured data are much less useful from an AI visibility perspective. Both matter, but your on-site schema should reflect your review standing accurately.