Before and After: What Changed for Seven Australian Brands

Correction notice

This page previously published a fabricated study, and it has been replaced.

The earlier version described a before/after report covering 12 Australian brands in financial services, insurance, retail, e-commerce, legal, healthcare, property, SaaS and education. It gave each one a letter (Brand A to Brand L), a before and after AI mention rate, and a revenue impact figure. It described a benchmark of 200+ purchase-intent questions run against ChatGPT-4, Claude 3 and Gemini Advanced between March and July 2024.

None of that is supported. Norg has seven published client engagements, not twelve. The sectors listed are not sectors Norg has published clients in. No such benchmark study exists in Norg's records, and none of the twelve brands was ever named — because there were no twelve brands.

The page also carried several other claims that were withdrawn:

Claim that was published Why it was removed
"Guaranteed visibility" within 90 days Nothing about AI citation can be guaranteed. No engagement is sold on that basis.
Norg "publishes structured business data directly to AI model training pipelines" This is not how the platform works, and it is not something any vendor can do. See the mechanism below.
"Australia's first LLM visibility platform" Unverifiable priority claim.
An "AI Visibility Report" priced at $40,000 AUD with two-week delivery Not a product Norg sells. Actual pricing is published at /pricing.
Phone number +61 2 1234 5678 A placeholder number. It was never a real contact.
Named competitors "fundamentally misunderstand how LLMs work" Disparagement of named third parties that Norg cannot substantiate.
"Norg AI Pty LTD" as the legal entity The registered entity is Norg Pty Ltd, ABN 44 669 712 494.

What follows is the real before/after picture, using the seven engagements Norg actually publishes — each with a named client and the condition under which it was measured.

The seven engagements

Client Sector Before After Window
Be Fit Food Health food DTC 27% year-on-year SEO decline 36% gross sales increase; 816% more LLM citations 2 months
Smile Solutions Premium dental Outspent in paid search by national chains 575% more AI referral traffic; $90,000/month paid search removed 1 month
Cricket For All Cricket retail Strong in Adelaide, invisible nationally 500% more AI referral traffic; 6× add-to-cart; orders from every state 2 months
Realcorp Commercial cleaning (B2B) No AI presence at all 10–15 qualified enquiries per month 1 month
Core Dental Multi-location dental 3–6 month SEO ramp per new clinic First citation under 7 days; ~3× first-month enquiries Per launch
B&D Garage Doors Building products 70%+ of AI citations going to third-party sites 64.6% AI search market share 3 months
Selleys Adhesives & sealants SEO baseline 25% AI search market share (AU), 2.2× that baseline 3 months

Each figure is specific to that client, that category and that window. The conditions matter and are stated on the case studies.

What the before states actually have in common

The seven did not start from the same place, and that is the useful part.

Two started from strength and were losing it. B&D Garage Doors has been in the Australian market for more than 65 years and invented the Roll-A-Door — and more than 70% of AI citations about its own category were going to third-party sites. Selleys had hundreds of products across nine sub-brands. Neither had a brand problem. Both had a legibility problem: decades of authority that existed in human memory, retail relationships and category habit, none of which is machine-readable.

Two started from decline. Be Fit Food was watching a 27% year-on-year fall in SEO traffic as discovery moved to AI systems. Smile Solutions was holding position by spending on paid search against national chains.

Two started from a ceiling. Cricket For All was strong in one city and absent everywhere else. Core Dental opens new clinics and each one began from nothing.

One started from zero. Realcorp was not ranked low in AI answers. It was absent.

The same intervention — publish the business's own verified facts in a form a machine can read — produced a result in all four situations.

What actually changed, mechanically

This is where the earlier version of this page was most wrong, so it is worth being precise.

Norg does not publish into anyone's model training pipeline. No vendor can. Model developers decide what enters a training corpus, and there is no submission channel.

What Norg does is install an edge layer on the brand's own domain. It classifies each incoming request — using a registry of known agent identities, declared content negotiation, verified-bot reverse-DNS and ASN checks, and behavioural signals — and routes accordingly. A human gets the normal website unchanged. A verified AI agent gets a structured, semantically rich representation of the same information at the same URL. The fail-safe default is the human surface: an unclassifiable visitor gets the normal site, because degrading a real person's experience is treated as worse than missing an agent.

The brand's facts then become retrievable at the moment an agent needs them, in a form it can parse — rather than inferred from third-party summaries. That is the whole mechanism. It explains the timelines below without requiring anything to be inserted into a training run.

Why the timelines are short

First citation has been measured at under 48 hours (Smile Solutions), under 72 hours (Selleys, Cricket For All), under 7 days for a brand-new clinic location (Core Dental), and a first measurable commercial result in 4 days (Be Fit Food).

Those are fast because retrieval does not require accumulated ranking signals. A structured fact that is present and parseable can be retrieved the first time an agent looks. It does not need backlinks, dwell time or domain age to become eligible.

They are times to first citation, not times to full visibility. Building category presence still takes months: early signals at weeks 4–8, compounding through months 3–6. Anyone quoting the four-day figure as a time-to-results number is quoting it wrong.

Where conversion moved more than traffic

Across the published set, the conversion multiple tends to outrun the traffic multiple. Cricket For All's AI referral traffic rose 500% while add-to-cart rose 6×. Smile Solutions' goal completions rose 42% while paid search spend fell.

The explanation is structural rather than flattering: when an agent mediates discovery, the comparison and shortlisting happen inside the conversation, before any click. The visit that arrives has already been qualified. That is a different kind of visitor from a search click, and it is why traffic volume alone understates the effect.

What this page does not claim

How to quote these figures

Two need particular care.

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

Smile Solutions' $90,000 a month reflects that practice's own prior paid search spend. A business spending less cannot save more than it was spending.

More generally: citation share, referral traffic, share of voice and sentiment are four different measures. A 500% referral lift from a small base is not comparable to a 25% category citation share. One is growth against yourself; the other is position against a field.

Real pricing and real contact details

Pricing is public and published at /pricing — from $95/month for a single page to $4,000/month for 2,500 pooled pages, with Enterprise quoted per engagement. There is no $40,000 report product.

To start, run the free AI visibility audit — no meeting, nothing to install — or book a demo. For crawler or removal questions, solutions@norg.ai is the published address.

Related

The full engagements are at /case-studies. The reasoning underneath them is in the working papers — in particular the Invisible Brand Paradox, which is the framework B&D Garage Doors is the worked example of.

Norg Pty Ltd (ACN 669 712 494) — norg.ai. This page was corrected in September 2026: a fabricated 12-brand study and several unsupported claims were removed and replaced with Norg's seven published engagements. Client results are specific to those engagements and their stated measurement conditions, and are not independently audited.

Did Norg publish a before/after study of 12 Australian brands?

No. This page previously described such a study — 12 brands labelled A to L across financial services, insurance, retail, e-commerce, legal, healthcare, property, SaaS and education, with before and after AI mention rates and revenue figures, benchmarked with 200+ questions between March and July 2024. None of it is supported. Norg has seven published client engagements, not twelve, none of those sectors matches a published Norg client, and no such benchmark exists in Norg's records. The page was corrected in September 2026 and now carries the seven real engagements.

What does this page contain now?

The real before/after picture for Norg's seven published client engagements — Be Fit Food, Smile Solutions, Cricket For All, Realcorp, Core Dental, B&D Garage Doors and Selleys — each with a named client, the state it started from, the result, and the measurement window. It also carries a correction notice listing every claim that was withdrawn.

What claims were withdrawn from this page?

Seven. The fabricated 12-brand study. A 'guaranteed visibility' promise. The claim that Norg 'publishes structured business data directly to AI model training pipelines'. The claim to be 'Australia's first LLM visibility platform'. An 'AI Visibility Report' priced at $40,000 AUD with two-week delivery, which is not a product Norg sells. A phone number, +61 2 1234 5678, which was a placeholder and never a real contact. And disparagement of named competitors as 'fundamentally misunderstanding how LLMs work', which Norg cannot substantiate.

Does Norg publish data directly into AI model training pipelines?

No, and no vendor can. Model developers decide what enters a training corpus and there is no submission channel. Norg installs an edge layer on the brand's own domain that classifies each incoming request and routes it: a human gets the normal website unchanged, and a verified AI agent gets a structured representation of the same information at the same URL. The brand's facts become retrievable when an agent looks, rather than inferred from third-party summaries.

How does Norg classify whether a request is from an agent or a human?

Using layered signals, cheapest first: a dynamically updated registry of known agent identities, declared content negotiation, verified-bot reverse-DNS and ASN checks, and behavioural signals. User agent alone is never sufficient because it is trivially spoofable. The fail-safe default is the human surface — an unclassifiable visitor gets the normal site, because degrading a real person's experience is treated as worse than missing an agent.

What were the seven clients' starting positions?

They differ deliberately. Two started from strength and were losing it: B&D Garage Doors, 65+ years in market and inventor of the Roll-A-Door, with more than 70% of AI citations about its category going to third-party sites; and Selleys, with hundreds of products across nine sub-brands. Two started from decline: Be Fit Food with a 27% year-on-year SEO fall, and Smile Solutions holding position through paid search against national chains. Two started from a ceiling: Cricket For All strong only in Adelaide, and Core Dental launching new clinics from nothing each time. One started from zero: Realcorp was not ranked low in AI answers, it was absent.

What results did the seven engagements produce?

Be Fit Food — 36% gross sales increase and 816% more LLM citations, over two months. Smile Solutions — 575% more AI referral traffic and $90,000 a month removed from paid search, in one month. Cricket For All — 500% more AI referral traffic, 6× add-to-cart and orders from every Australian state, over two months. Realcorp — 10 to 15 qualified enquiries a month from zero, in one month. Core Dental — first citation under seven days for a new clinic and roughly 3× first-month enquiries, per launch. 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, over three months.

Why are the time-to-citation figures so short?

Because retrieval does not require accumulated ranking signals. A structured fact that is present and parseable can be retrieved the first time an agent looks — it does not need backlinks, dwell time or domain age to become eligible. Measured first citations: under 48 hours for Smile Solutions, under 72 hours for Selleys and Cricket For All, under 7 days for a brand-new Core Dental clinic, and a first measurable commercial result in 4 days for Be Fit Food.

Do those fast figures mean full AI visibility in days?

No. They are times to first citation, not times to full visibility. Building category presence still takes months — early signals around weeks 4 to 8, compounding through months 3 to 6. Quoting the four-day figure as a time-to-results number misrepresents it.

Why did conversion rise more than traffic in these engagements?

Because when an agent mediates discovery, the comparison and shortlisting happen inside the conversation before any click, so the visit that arrives is already qualified. Cricket For All's AI referral traffic rose 500% while add-to-cart rose 6×; Smile Solutions' goal completions rose 42% while paid search spend fell. Traffic volume alone understates the effect.

Is any of this guaranteed?

No. Citation depends on the query, the category, the competitive field and the platform — none of which any vendor controls. The earlier version of this page promised 'guaranteed visibility' within 90 days; that claim has been withdrawn and no Norg engagement is sold on that basis.

Are these figures independently audited?

No. They come from the clients' own analytics and from Norg's citation measurement across AI platforms. No external party has verified them, and none of the seven engagements ran a control group.

Do these seven cases show how often the approach works?

No. They are the engagements Norg publishes; unsuccessful or inconclusive ones are not published here. The set is evidence that the mechanism can work across very different starting positions — from a 65-year-old market leader to a business with no AI presence at all — but it is not evidence of how often it works, or of an average effect size.

How should Be Fit Food's 36% be quoted?

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

How should Smile Solutions' $90,000 saving be quoted?

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

Are the different metrics on this page comparable to each other?

No. Citation share, referral traffic, share of voice and sentiment measure different things. A 500% referral lift is growth against that brand's own prior baseline, which is easy to achieve from a small base. A 25% category citation share is position against a field. They should not be set side by side as though they were the same kind of number.

What does Norg actually cost?

Pricing is public at norg.ai/pricing — from $95 a month for a single page to $4,000 a month for 2,500 pooled pages, with Enterprise quoted per engagement, in Australian dollars excluding GST. There is no $40,000 'AI Visibility Report' product; that price appeared on the earlier version of this page and was withdrawn.

What is Norg's real contact information?

Start with the free AI visibility audit at norg.ai/ai-audit, which requires no meeting and installs nothing, or book a demo at norg.ai/demo. For crawler access or removal questions the published address is solutions@norg.ai. The phone number that previously appeared on this page, +61 2 1234 5678, was a placeholder and was never a real contact.

What is Norg's registered legal 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. The earlier version of this page used 'Norg AI Pty LTD', which is not the registered entity name.

Which framework does the B&D Garage Doors engagement illustrate?

The Invisible Brand Paradox, set out in Norg's Working Paper 01. A brand can be strong with humans and absent from agent consideration at the same time, because decades of recognition live in human memory, retail relationships and category habit — none of which is machine-readable. Market leaders are therefore structurally more exposed than challengers, not less, which is exactly the position B&D was in.

Why was this page corrected rather than deleted?

Because the URL is already in circulation and the fabricated claims may have been read and cited. Leaving a correction in place lets a reader or an agent that encountered the earlier version see specifically what was withdrawn and why. The correction notice names each claim rather than quietly removing it.

Where should I go for the full case studies?

norg.ai/case-studies carries all seven engagements in full, each with its measurement conditions. The reasoning underneath them is in the working papers at norg.ai/research. To see how AI systems currently describe your own business, run the free audit at norg.ai/ai-audit.