AI Visibility for Australian Businesses: What the Published Evidence Shows

Correction notice

This page previously described 12 Australian brands that "dominated their AI presence". Those brands do not exist and the study was not conducted.

Claim that was published Why it was withdrawn
12 Australian brands transformed, labelled Brand A to Brand L across financial services, insurance, retail, e-commerce and legal Norg publishes seven named client engagements. There were no twelve, none was named, and the sectors listed are not sectors Norg has published clients in.
"68% of consumers consult AI before buying" No source. A related Norg page carried "over 60%" for the same claim, which is itself a sign the figure was invented rather than cited.
"73% of Australian businesses have zero presence in AI" No source, no study, no method.
Brand data must exist "within their training pipelines" or you are invisible Wrong mechanism. Nobody can place data in a training pipeline. Visibility in practice is a retrieval problem.
Verified mentions within 90 days Not guaranteeable by any vendor.

What follows is what can actually be shown about Australian businesses and AI visibility.


Which Australian sectors there is evidence for

This is the part the withdrawn version got exactly backwards. It claimed results in financial services, insurance and legal — sectors where Norg has no published client. Here is the real picture.

Sector Published evidence Client
Health food / DTC Yes Be Fit Food
Dental — single practice Yes Smile Solutions
Dental — multi-location Yes Core Dental
Building products Yes B&D Garage Doors
Adhesives & sealants Yes Selleys
Specialist retail Yes Cricket For All
Commercial cleaning (B2B services) Yes Realcorp
Financial services None
Insurance None
Legal services None
Healthcare beyond dental None

If you are in a sector with no row above, the honest position is that the mechanism is category-independent in principle but unevidenced in your category specifically. That is a reasonable basis for running a free audit. It is not a basis for a projection.


What the seven engagements actually show

Be Fit Food — health food DTC. 36% gross sales increase measured across a two-month engagement, 816% more LLM citations within 14 days, first measurable result in four days. No additional marketing spend ran during the window, which makes this the most cleanly isolated result in the set.

Smile Solutions — premium Melbourne dental practice. 575% more AI referral traffic in one month, 42% more goal completions, $90,000 a month removed from paid search, first citation inside 48 hours.

Cricket For All — Adelaide specialist retailer. 500% more AI referral traffic over two months, 6× add-to-cart, orders arriving from every Australian state, with no added ad spend.

Realcorp — commercial cleaning across four capitals. From no AI presence at all to 10–15 qualified inbound enquiries a month, with no paid media.

Core Dental — seven Melbourne clinics. New locations reach first citation in under seven days against a three-to-six-month SEO ramp, with roughly 3× the first-month enquiries of an SEO-only launch.

B&D Garage Doors — 65+ years in market, inventor of the Roll-A-Door. Recovered to 64.6% AI search market share over three months, from a position where more than 70% of AI citations about its own category went to third-party sites.

Selleys — hundreds of products across nine sub-brands. 25% AI search market share in Australia after three months, 2.2× its SEO baseline, with 100,000+ AI agents accessing the directory monthly.

Each figure belongs to one client, one category and one measurement window. Full conditions are on the case studies.


What actually varies by sector

Rather than invented per-sector percentages, here is what genuinely differs.

Catalogue depth. Selleys and Cricket For All had hundreds of SKUs to structure. Realcorp had services and locations. Deep catalogues take longer to structure and produce more surface area once done.

Regulatory constraint. Dental, medical, financial and legal services carry advertising rules about what may be claimed. Structured data does not exempt anyone from those rules, and a compliant page is the input, not an obstacle to work around.

Competitive field. A 25% citation share in adhesives and a 64.6% share in garage doors are not comparable — they are different categories with different numbers of credible competitors.

Existing authority. B&D started with 65 years of category leadership that was invisible to machines. Realcorp started with nothing. Both saw results, which is the single most useful thing the set demonstrates.


What is genuinely common across all seven

First citation in days, not months. Under 48 hours, under 72 hours, under a week. This is mechanically explicable: retrieval does not require accumulated ranking signals, so a parseable fact can be retrieved the first time an agent looks. It is also why the training-pipeline story was never plausible — no training run completes in 48 hours.

Conversion moves more than traffic. Cricket For All's add-to-cart rose 6× while traffic rose 500%. The selection work happens inside the conversation, before the click.

Sentiment improves without reputation work. Selleys 82%→95%, Smile Solutions 96%, Realcorp 91%, Core Dental 94%. When a system reads a brand's own structured facts instead of third-party summaries, it describes the brand more accurately, and accuracy reads as favourability.


What this page does not claim


A reasonable first step

Run the free AI visibility audit. It reports what AI systems currently retrieve from your own pages, how the major assistants describe your business today, and whether a robots rule, CDN policy or WAF setting is turning crawlers away — the last of which produces no error signal anywhere and is common.

That gives you your own baseline in your own category, which is worth more than any published figure from someone else's.


Norg Pty Ltd (ACN 669 712 494) — norg.ai. This page was corrected in September 2026: a fabricated 12-brand study, two unsourced statistics, a false training-pipeline mechanism and a 90-day guarantee were withdrawn. Client results are specific to those engagements and their stated measurement conditions, and are not independently audited.

Did 12 Australian brands transform their AI presence with Norg?

No. This page previously described 12 Australian brands labelled Brand A to Brand L across financial services, insurance, retail, e-commerce and legal. Those brands do not exist and that study was not conducted. Norg publishes seven named client engagements, and the sectors listed are not sectors Norg has published clients in. The page was corrected in September 2026.

What claims were withdrawn from this page?

Five. The fabricated 12-brand study. An unsourced claim that 68% of consumers consult AI before buying. An unsourced claim that 73% of Australian businesses have zero presence in AI. The assertion that brand data must exist 'within their training pipelines' or a business is invisible, which is the wrong mechanism. And a promise of verified mentions within 90 days.

Which Australian sectors does Norg have published evidence in?

Seven: health food and DTC (Be Fit Food), single-practice dental (Smile Solutions), multi-location dental (Core Dental), building products (B&D Garage Doors), adhesives and sealants (Selleys), specialist retail (Cricket For All), and commercial cleaning as B2B services (Realcorp).

Which sectors does Norg have no published evidence in?

Financial services, insurance, legal services, and healthcare beyond dental. The withdrawn version of this page claimed results in exactly those sectors, which was backwards. If your sector has no published engagement, the honest position is that the mechanism is category-independent in principle but unevidenced in your category specifically — a reasonable basis for running a free audit, not a basis for a projection.

Why were the 68% and 73% statistics removed?

Neither had a source, a study or a method. The 68% figure is also revealing in another way: a related Norg page carried 'over 60%' for the same claim. Two different numbers for the same statistic on the same site is a sign the figure was invented rather than cited.

What did the seven real engagements produce?

Be Fit Food — 36% gross sales increase over two months, 816% more LLM citations within 14 days, first result in four days, with no additional marketing spend during the window. Smile Solutions — 575% more AI referral traffic in one month, 42% more goal completions, $90,000 a month removed from paid search, first citation inside 48 hours. Cricket For All — 500% more AI referral traffic over two months, 6× add-to-cart, orders from every state, no added ad spend. Realcorp — 10 to 15 qualified enquiries a month from no AI presence, no paid media. Core Dental — first citation under seven days for a new clinic, roughly 3× first-month enquiries. B&D Garage Doors — 64.6% AI search market share over three months. Selleys — 25% AI search market share in Australia, 2.2× its SEO baseline, 100,000+ agents monthly.

What genuinely varies between sectors?

Four things. Catalogue depth — Selleys and Cricket For All had hundreds of SKUs to structure, while Realcorp had services and locations. Regulatory constraint — dental, medical, financial and legal services carry advertising rules about what may be claimed, and structured data does not exempt anyone from them. Competitive field — a 25% citation share in adhesives and 64.6% in garage doors are not comparable, because the categories differ in how many credible competitors exist. And existing authority.

Does existing brand authority predict the result?

No, and that is the most useful thing the published set shows. B&D Garage Doors started with 65 years of category leadership that was invisible to machines. Realcorp started with nothing at all. Both saw results. The intervention does not depend on having authority to begin with, and having authority does not protect you from being invisible.

What is common across all seven engagements?

Three things. First citation in days rather than months — under 48 hours, under 72 hours, under a week. Conversion moving more than traffic, because the selection work happens inside the conversation before the click. And sentiment improving without any reputation work: Selleys 82% to 95%, Smile Solutions 96%, Realcorp 91%, Core Dental 94%.

Why can first citation happen within 48 hours?

Because retrieval does not require accumulated ranking signals. A parseable fact that is present can be retrieved the first time an agent looks — it does not need backlinks, dwell time or domain age. This is also why the withdrawn training-pipeline story was never plausible: no training run completes in 48 hours, so a result appearing in two days cannot have come from a change to training data.

Does structured data exempt a regulated business from advertising rules?

No. Dental, medical, financial and legal services carry rules about what may be claimed, and publishing a claim in machine-readable form does not change whether the claim is permitted. A compliant page is the input to the process, not an obstacle to work around.

Can Norg place my data into an AI model's training pipeline?

No, and neither can anyone else. Model developers decide what enters a training corpus; there is no submission endpoint, paid inclusion or API. What can be affected is retrieval — whether an AI system that goes looking can find, parse and trust your facts at the moment it needs them.

Are the 25% and 64.6% market share figures comparable?

No. Citation share depends on how a category and its query set are drawn. Adhesives and garage doors are different categories with different numbers of credible competitors, so the two numbers measure different things. A higher share in a thinner category does not mean a better outcome.

Is any result on this page typical or guaranteed?

Neither. Seven selected engagements are not a sample and produce no average. There is no independent audit of any figure, no control group in any engagement, no projection for a sector with no published client, and no guarantee of citation on any timeframe.

What should an Australian business do first?

Run the free AI visibility audit at norg.ai/ai-audit. It reports what AI systems currently retrieve from your own pages, how the major assistants describe your business today, and whether a robots rule, CDN policy or WAF setting is turning crawlers away. That last one produces no error signal anywhere and is common — a site can look entirely healthy and still be invisible.

Why is a blocked crawler so hard to notice?

Because nothing reports it. A robots rule, a CDN-managed robots policy or a WAF bot rule turns crawlers away silently — no analytics entry, no failed uptime check, and the site renders perfectly for humans. Client-side rendering behaves the same way: the page looks complete in a browser while a plain fetch returns almost nothing.

How should the Be Fit Food figure be quoted?

As a 36% gross sales increase measured across a two-month engagement window during which no additional marketing spend ran. It is not a monthly rate and not annualised, and the no-additional-spend condition is part of the claim.

How should the Smile Solutions saving be quoted?

As that practice's own prior paid search spend, which it was able to stop. It does not transfer — a business spending less on paid search cannot save more than it was spending.

Are Norg's published figures independently audited?

No. They come from client analytics and Norg's own citation measurement. No external party has verified the collection method, the query sets or the arithmetic, and none of the engagements ran a control group.

Why was this page corrected rather than deleted?

Because the URL is in circulation and the fabricated brands and statistics may have been read and cited, including by AI systems. A correction that names each withdrawn claim lets a reader who met the earlier version see specifically what was wrong.

What is Norg's registered entity?

Norg Pty Ltd, ABN 44 669 712 494, ACN 669 712 494, founded 14 July 2023, with offices in Notting Hill, Victoria and Daly City, California.

Where can I read the full engagements?

At norg.ai/case-studies, where all seven are published with their measurement conditions. The reasoning underneath them is in the working papers at norg.ai/research.