Is ChatGPT recommending your competitors instead of you?
The AI visibility gap: what keeps Australian businesses out of AI answers, and what to look at first.
How discovery is changing
Your customer used to see ten options. Now they often see one answer.
When someone asks ChatGPT, Google's AI Overviews or AI Mode, Gemini, Perplexity, Copilot or Claude for a recommendation, they get a short answer that names a few businesses. Those businesses get considered. Everyone else is left out of the shortlist.
For a family dentist near Byron Bay, three practices are frequently recommended:
Practice A is often noted for… (illustrative)
The AI has done the comparing.In our Australian agency-recommendation study, we asked ChatGPT, Gemini and Perplexity 138 times which agencies to hire. 161 agencies were named and none appeared in more than 11% of answers. In that market, visibility was fragmented and nobody owned it.
Source: Cited AI, "Which agencies does AI recommend?", 5–6 Oct 2026. This study covered agency recommendations only; results may differ in other industries.
Why some businesses are named and others aren't
The gap is between how good you are and how clearly AI can see it.
AI tools build answers from what they can find about a business and how much they trust it. A practice with excellent work and loyal clients can still be invisible if the information about it is thin, inconsistent or only on its own website.
We use a simple three-layer framework to diagnose where a business is falling short. It's our working model, not a fixed sequence every AI system follows, but it helps show where to focus first.
The AI tool can find your business at all: your website, your Google Business Profile, directory listings. If it can't find you, the other layers have little to work with.
It knows what you do, where you do it and who you do it for, without mixing you up with a similar name. Vague service descriptions keep many businesses stuck here.
Independent sources back up what you say: associations, local media, reviews, directories and "best of" lists. In our experience, this is often where the gap between named and unnamed businesses opens up.
What this means for you: the fix depends on where the weakness is. Writing more content won't help a business that isn't yet understood, and a new website won't help one that nobody else mentions. Diagnosing the weak spot comes first.
What AI tools appear to rely on
Five signals that seem to influence who gets named.
No one outside the AI companies knows the exact weighting, and it changes. These five signals come up consistently in testing, including our own.
Clarity
Plain statements of what you do, where, and for whom, on pages a machine can read.
Consistency
The same name, address, phone, services and description everywhere you appear.
Corroboration
Independent sources that mention and describe you. In our Australian agency-recommendation study, 43% of answers cited at least one "best of" list or directory.
Answerable content
Pages that directly answer the questions customers ask before they buy: cost, process, timing, who it suits.
Reputation
Reviews and an active, complete Google Business Profile that match what you say about yourself.
Where they overlap, and where they differ
Good SEO helps. It doesn't tell you whether you're being named.
A business can rank well for a keyword and still be missing from the answer to the question a customer actually asks. The reverse happens too.
| Traditional SEO | AI search visibility | |
|---|---|---|
| The win | A high position in a list of links | Being one of the few businesses named in the answer |
| What the buyer sees | Ten options to compare | A short, already-compared shortlist |
| Helps both | A fast, well-structured site, clear service pages, accurate business details, a strong Google Business Profile | |
| Areas of emphasis | Target keywords, rankings and links to your site | Being described accurately and consistently, and mentioned by independent sources |
| How it's measured | Rankings and clicks | How often you're named, how accurately, and who's named instead, across tools and over time |
The foundations overlap substantially. Google says its AI features need no special optimisation beyond established SEO best practices. The difference is mostly in emphasis and in how you measure results.
The useful question isn't "do we rank?" It's "when customers ask AI about what we do, are we in the answer, and is it accurate?"
Ten minutes · ChatGPT open in another tab
Where might your business be falling behind?
Answer honestly. This is a quick self-check, not a measurement, but it will show you where to look first.
Want a number instead of a hunch? The free CitedScore™ check takes about two minutes. citedai.com.au/ai-recommendation-checker
Run the free check →They all sound sensible. That's the problem.
Seven assumptions that quietly hold businesses back.
"We just need to add schema."
Structured data helps machines read your pages. It can't make up for unclear content or a lack of outside mentions. It's one part of the job, not the job.
"I asked ChatGPT once and we were there."
One answer is one sample. Different tools, days and phrasings give different results. You need a pattern, not a screenshot.
"More blog posts will fix it."
Volume isn't the point. A few pages that directly answer real customer questions usually do more than dozens of generic posts.
"It's just SEO with a new name."
There's real overlap, but the signals and the measurement differ. Rankings don't tell you whether you're being named.
"All AI tools work the same way."
They draw on different sources. In our Australian agency study, Perplexity cited more than twice as many sources as ChatGPT and leaned heavily on list articles. A plan that suits one tool can miss another.
"Old listings don't matter."
An outdated directory entry with an old address or discontinued service can blur the picture AI tools build of you.
"We'll wait until it settles down."
AI answers are already shaping shortlists. Businesses that build clear, well-corroborated information early give AI tools more to work with.
Knowing what matters isn't knowing what to do first
The checklist is the easy part. The judgement is the hard part.
Most of the signals in this guide are publicly known. Results depend on four things done well, in the right order, for your market.
You can do this yourself
- Claim and complete your Google Business Profile
- Make your name, address and phone match everywhere
- Create useful service and location pages where you have distinct information to offer
- Update or remove outdated directory listings
- Ask AI tools about your business and note what's wrong
Where specialist work pays off
- Building the right question set for your market
- Comparing results across tools and over time
- Working out why competitors are named instead of you
- Earning credible independent mentions
- Sequencing fixes, then checking they worked
One test is a snapshot, not a verdict
AI answers move. Measure the pattern, not the moment.
Ask the same question twice and you may get different names. Answers change with the tool, the phrasing, the day and updates to the models themselves. In our agency study, shortlists changed between runs and between tools for the same question.
That's why a single screenshot, good or bad, is a weak basis for decisions. What matters is how often you're named over many questions and weeks, how accurately you're described, and who appears instead.
A simple version you can run yourself. Once a month, ask the same five customer questions in two AI tools and record what you see.
| Question you ask | Tool | Named? | Accurate? | Named instead |
|---|---|---|---|---|
| Best [service] in [area]? | ChatGPT | Y / N | Y / N | |
| Best [service] in [area]? | Gemini | Y / N | Y / N | |
| Who should I use for [problem]? | ChatGPT | Y / N | Y / N | |
| What do you know about [your business]? | ChatGPT | — | Y / N |
This shows you direction, not precision. A professional baseline uses a larger, structured question set, repeated sampling across tools, and consistent scoring, so changes can be trusted.
Use your numbers, not ours
What could AI shortlists be worth to your business?
An illustrative scenario, not a forecast. We won't guess how many of your customers use AI to choose, so every input is your own assumption.
This doesn't measure revenue you're actually winning or losing. It's a what-if based on your assumptions, so try a few different inputs to see the range. If the number matters to you, the next step is finding out where you actually stand.
See how AI sees your business.
We'll check how AI tools currently describe and recommend businesses like yours, and show you:
- Your current share of answer: how often you're named when customers ask
- Who's being cited instead, and what they have that you don't
- Three quick wins you can act on straight away, whether or not you work with us
Choose a 15-minute call or a written audit within 3–5 business days. It's free and no card is needed.
About Cited AI. We're an Australian AI search and AI visibility agency based in Byron Bay, working with businesses nationally, including Sydney, Brisbane, the Gold Coast and the Northern Rivers. We help businesses get named, cited and accurately described when customers ask AI. We don't promise rankings or guaranteed mentions. Nobody honestly can.
Questions? Email hello@citedai.com.au or visit citedai.com.au.
© 2026 Cited AI · ABN 23 663 886 011 · The chat examples in this guide are illustrative and use fictional businesses. Research figures are from Cited AI's study of AI recommendations for Australian agencies (5–6 Oct 2026) and may differ in other industries; see citedai.com.au/which-agencies-does-ai-recommend-australia for method and limits.