Most Australian ecommerce founders have never run this test. It takes five minutes, costs nothing, and changes what they do next. This guide gives you the exact prompts, a scoring system for the results, and the one column in your notes that matters more than whether you appear.
Here is the scale of it. Adobe Analytics tracked over one trillion visits to US retail sites and found AI-referred traffic grew 393% year on year in the first quarter of 2026 (Adobe, April 2026). In March 2026 that traffic converted 42% better than every other channel, a reversal from converting 38% worse a year earlier.
KEY FACTS
- AI-referred traffic to US retail sites grew 393% year on year in Q1 2026 (Source: Adobe Analytics, April 2026)
- AI traffic converted 42% better than non-AI traffic in March 2026, after converting 38% worse a year earlier (Source: Adobe Analytics, April 2026)
- AI Overviews appeared in 25.11% of 21.9 million Google searches analysed (Source: Conductor 2026 AEO/GEO Benchmarks Report)
- 87.4% of all AI referral traffic comes from ChatGPT (Source: Conductor, 3.3 billion sessions across 13,770 domains)
- AI referrals account for 1.08% of all website traffic on average (Source: Conductor 2026 Benchmarks)
- The average retail product page is 66% readable by large language models (Source: Adobe AI Content Visibility Checker, April 2026)
- Australians spent $82.6 billion online in 2025, up 14% year on year (Source: Australia Post eCommerce Report 2026)
What exactly should I type into ChatGPT?
Type the question your customer asks, not your brand name. Searching your own name proves nothing, because the engine finds you either way. The test is whether you get named when nobody mentions you first. Write five buyer questions as full sentences, not keywords, and run each one in four engines.
Use this pattern. Replace the brackets with your own category.
The category question. "What are the best organic skincare brands in Australia?"
The problem question. "I have sensitive skin and live in Melbourne. What moisturiser should I buy?"
The comparison question. "Is [competitor] or [competitor] better for daily use?"
The price question. "Where can I buy protein powder online in Australia for under $60?"
The trust question. "Which Australian [your category] brands ship fastest?"
Open a fresh chat for each prompt. An existing chat remembers what you discussed earlier, and that memory puts your brand into answers a stranger never sees. A fresh chat is the honest test.
Record three things every time: whether your brand appears, which brands appear, and which websites the engine cited as sources. The third column is the one almost nobody writes down, and it is the one that tells you what to do next.
Do not have five minutes? We run the full 20-prompt version live on a call and show you the answers on screen. Free, 15 minutes, no obligation. Book a free AI visibility call.
Which AI engines should I test, and does the order matter?
Test ChatGPT first, because it dominates the referrals. Conductor analysed 3.3 billion sessions across 13,770 domains and found 87.4% of all AI referral traffic comes from ChatGPT. Then test Gemini, Perplexity and Claude. Add Google AI Overviews by running your five questions in normal Google search and reading what appears above the blue links. This is the same panel we run inside our ecommerce SEO and AI search programme.
Each engine draws on different sources and refreshes at a different speed. That is why the same question returns four different answers.
| Engine | Leans most on | How fast it updates | Counted in the GA4 AI Assistant channel |
|---|---|---|---|
| ChatGPT | Authoritative editorial, news, reference sites | Roughly weekly | Yes |
| Google AI Overviews and AI Mode | Reddit, YouTube, Quora, LinkedIn, news | Continuous | No. Filed as Organic Search |
| Perplexity | Reddit, video transcripts, news | Real time | No. Lands in Referral |
| Gemini | Google's index | Continuous | Yes |
| Claude | Editorial and academic sources | Real time | Named at launch, list has since changed |
Google AI Overviews deserve their own slot because of how common they now are. Conductor analysed 21.9 million Google searches between 15 September and 12 October 2025 and found 25.11% returned an AI Overview. In health care that figure reached 48.7%.
Do not treat any single number as settled. Prevalence studies range from about 16% to over 60% depending on the query set and the method. The direction is what matters, and the direction is up.
What does a good result look like versus a bad one?
A good result names your brand in the first answer, unprompted, and links your own site as a source. A bad result names three competitors and cites a list post you have never heard of. Most Australian ecommerce brands get the bad result on all five questions the first time they run this test. That is the normal starting point, not a disaster.
Score every prompt one of four ways.
| Result | What you saw | What it means | What to do first |
|---|---|---|---|
| Named and cited | Your brand appears and your site is linked as a source | The engine trusts your own pages | Protect it. Do not restructure those pages |
| Named only | Your brand appears, a third-party page is the source | Other people's writing carries you | Find that page and strengthen the relationship |
| Missing, category open | Competitors named, none dominant | The gap is winnable | Start with your own page structure |
| Missing, category owned | One brand named in every answer | Slow and expensive to shift | Fix conversion first, then compete |
The gap between "named" and "cited" is the one worth understanding. A mention puts your brand in the sentence. A citation puts your URL in the answer as a source. Mentions come from what other websites say about you. Citations come from your own pages being readable and quotable.
Now read your source column. If the same three or four websites appear across all 20 prompts, those sites are the gatekeepers in your category. Getting listed on one of them moves your visibility faster than anything you publish yourself.
Want us to run this on your category? We do the 20-prompt version live on a call, across all four engines, and show you the answers on screen. Free, 15 minutes, no obligation. Book a free AI visibility call.
Why does one engine name my brand and another does not?
Because each engine trusts a different corner of the internet. Perplexity weights Reddit and video transcripts heavily. Claude favours editorial and academic sources. Google AI Overviews pull from Reddit, YouTube and Quora. A brand with strong Google rankings surfaces in Gemini first, because Gemini works from Google's index.
Three patterns explain almost every inconsistency you will see.
Freshness differs. ChatGPT reindexes roughly weekly. Perplexity and Claude retrieve in real time. A page you published last week reaches one engine before another.
Source weighting differs. If your category is discussed heavily on Reddit, Perplexity finds you and Claude does not. If your category is covered by trade publications, the reverse happens.
Your entity signals differ by engine. Engines confirm a brand exists by finding the same details in several places: your site, your Google Business Profile, your social profiles, third-party listings. Thin or inconsistent details make you easy for one engine to resolve and hard for another.
None of that is fixable in a week. All of it is fixable in a quarter.
How do I check the AI traffic already arriving in Google Analytics?
Open Google Analytics 4, then Reports, then Acquisition, then Traffic acquisition. Look for a channel called AI Assistant. Google added it on 13 May 2026 and it requires no setup. Before that date, a click from ChatGPT looked identical to a link from any random website, because both landed in Referral.
Three things to know before you read the number.
- It counts forward only. Sessions before 13 May 2026 stay in Referral. Your history is not restated.
- Perplexity is not in it. Google's earlier guidance for building a custom channel named ChatGPT, Gemini, Microsoft Copilot, Claude and Perplexity. The default channel names a smaller set, and that set changed once since launch. Check it on the day you report.
- Referrer-free clicks land in Direct. Anyone arriving from an AI app that strips the referrer counts as Direct, not AI.
Every AI number in your analytics is a floor, never a ceiling.
How much AI traffic should I expect right now?
Expect a very small number. Conductor measured AI referrals at 1.08% of all website traffic across 3.3 billion sessions and 13,770 domains. For most Australian ecommerce stores the figure sits under 2%. A result in that range is normal in August 2026, not a sign that something is broken.
The argument for acting now is not the current volume. It is the combination of 393% annual growth and traffic that converts 42% better than everything else once it arrives. A channel is cheapest to enter before it gets crowded.
One warning before you calculate your own percentage. Split your traffic by country and engagement rate first. Automated traffic inflates session counts on small sites and drags every percentage out of shape. If a large block of sessions leaves within a second of arriving, strip it out before you divide anything.
What do I do if my brand is not there at all?
Do not start writing blog posts. Make the pages you already have readable and quotable first, then work on getting named on the third-party pages your test showed the engines citing. The Princeton GEO study (Aggarwal and colleagues, KDD 2024) tested nine techniques and found adding inline citations lifted AI citation rates by 40% and adding specific statistics lifted them by 35%.
That same study found five techniques did nothing at all, including keyword stuffing and padding out word counts. One actively hurt: marketing language with no substance behind it. Pages making claims with nothing to back them get filtered out of AI answers entirely.
Work in this order.
Confirm a crawler can read your pages. Adobe scored the average retail product page at 66% machine-readable, meaning a third of the content is invisible to a language model. Open your own product page, right click, choose View Page Source, and search for your product description. If it is not in the source, no engine sees it.
Put the answer at the top. Every page that matters answers its own question in the first 60 words, plainly, with a fact in it.
Give engines something worth quoting. "Fast shipping" is not quotable. "Dispatched from Melbourne within 24 hours" is.
Make your brand details match everywhere. Same name, same address, same founder across your site, your Google Business Profile and your social profiles.
Get onto the pages your test surfaced. Those sites are already quoted about your category. Being listed on one is faster than outranking anybody.
Not sure which of the five applies to you? The free audit walks through all five against your actual store and tells you which one is costing you most. Book the 15-minute AI visibility call.
How often should I re-run this test?
Run it once a month, on the same five questions, on roughly the same date. Monthly catches a real shift and ignores daily noise. Keep every result in one sheet so you read a trend rather than a snapshot. Expect nothing to move for 8 to 12 weeks, because new content takes that long to enter live retrieval indexes.
Four columns is all you need.
| Column | What goes in it | Why it matters |
|---|---|---|
| Date and engine | 27 Aug 2026, ChatGPT | Lets you compare like with like month on month |
| Named | Yes or no | The headline score |
| Who else appeared | Three competitor names | Shows whether the category is opening or closing |
| Sources cited | Three URLs | Tells you where to spend next month's effort |
Anyone promising you visibility in 30 days is describing something other than sustainable organic growth. Technical fixes register within weeks. Citation frequency and authority build across quarters.
Who should not bother with this yet?
Three types of Australian ecommerce brand should fix something else first. This is the section that costs us work, and it is here because sending you down the wrong path costs you more.
- You need revenue this month. AI visibility compounds over quarters. Email and SMS retention moves faster.
- Your store converts poorly. Sending more traffic to a leaky store makes the leak bigger. Fix the website first.
- You will not publish content. Citability requires something worth citing. If nothing is going to get published, the ceiling is low.
If none of those apply, run the test this week.
Frequently asked questions
Does asking ChatGPT about my own brand help it remember me?
No. ChatGPT does not learn from your conversations in a way that changes what it tells other people. Asking about yourself inside a chat where you already mentioned your brand makes you appear when a stranger would not see you. That is why every test prompt starts in a fresh chat.
Why does my brand appear in Gemini but not ChatGPT?
The two engines draw on different sources and refresh at different speeds. ChatGPT leans on authoritative editorial and reference sites and reindexes roughly weekly. Gemini works from Google's index and updates continuously. A brand with strong Google rankings surfaces in Gemini first.
Is being mentioned the same as being cited?
No. A mention is your brand name appearing in the answer text. A citation is your URL linked as a source. Mentions come mostly from what other websites say about you. Citations come from your own pages being readable, structured and quotable. Both count, and they need different work.
How many AI engines do I actually need to test?
Four covers the ground: ChatGPT, Gemini, Perplexity and Claude. Conductor found 87.4% of AI referral traffic comes from ChatGPT, so start there. Add Google AI Overviews by running the same questions in normal Google search and reading what sits above the blue links.
Does AI search traffic actually convert?
Adobe found AI-referred traffic to US retail sites converted 42% better than non-AI traffic in March 2026, reversing from 38% worse a year earlier. That is Adobe's own analytics data from over a trillion US retail visits. Treat it as directional for an Australian store rather than a benchmark.
How much of my traffic should come from AI right now?
Around 1%. Conductor measured AI referrals at 1.08% of all website traffic across 3.3 billion sessions and 13,770 domains. Most Australian ecommerce stores sit under 2%. The case for acting now rests on 393% annual growth and higher conversion, not on current volume.
Why is my GA4 AI Assistant number zero?
Three common reasons. The channel counts forward only from 13 May 2026. It excludes Google AI Overviews, which are filed as Organic Search. And clicks from AI apps that strip the referrer land in Direct. A zero means "not measured", not "nobody came".
Do I need a paid AI visibility tracking tool?
Not to start. Twenty prompts in a spreadsheet once a month gives you the trend, the competitor set and the source list, which is most of the value. Paid tools earn their keep once you track dozens of queries across multiple brands or need daily monitoring.
How long before changes to my site show up in AI answers?
Expect 8 to 12 weeks. Technical fixes such as making content readable in the page source register sooner. Citation frequency, entity recognition and topical authority build across quarters. Anyone promising 30 days is describing something else.
Should I block AI crawlers from my site?
Not if you sell online. Blocking OAI-SearchBot, PerplexityBot, ClaudeBot or Google-Extended removes you from the answers those engines give. Check your robots.txt file at yourdomain.com/robots.txt before anything else, because some website builders block them by default.
What if a competitor is named in every single answer?
Read their citations rather than their website. Note which third-party pages the engines quote when they name that competitor. Those pages are the gatekeepers. Getting listed alongside them moves you faster than trying to outrank the competitor directly.
Does schema markup get me into AI answers?
Schema gives engines a machine-readable summary of what a page is about and what facts it contains. It is not a direct ranking factor, and it removes ambiguity at the moment an engine decides whether it understands your page well enough to cite it. Product, FAQ, Article and Organization matter most for ecommerce.
Run it this week
You cannot fix a gap you have not measured, and the measurement is free. Twenty prompts, four engines, one spreadsheet, five minutes. Keep the source column, repeat it monthly, and you will know more about your AI visibility than most of your competitors know about theirs. If you would rather watch someone run it live on your own category, book a free AI visibility call and we will do all 20 prompts while you watch.
Ewen Ling, Founder of Elinix AI
Ewen Ling is the Founder of Elinix AI and has over 16 years of ecommerce industry experience across website conversion optimisation, email and SMS marketing automation, retention marketing, and ecommerce growth strategy. Elinix AI helps Australian ecommerce brands fix revenue leaks, improve conversion, and build sustainable customer retention systems.
**
