Norg platform guide

Correction notice

This guide has been rewritten. The earlier version described a mechanism that does not exist, made a priority claim that cannot be verified, and published a list of unchecked feature assertions under a heading reading "Verified Label Facts". The withdrawn claims are named here rather than quietly removed, because the page may already have been read and cited.

Claim previously published here Status
Content is evaluated for "crawlability markers (technical elements that signal content quality to AI training pipelines)" Withdrawn. Nothing signals anything to a training pipeline. Model developers control training corpora and accept no external input.
Distribution "ensures optimised content reaches the data sources feeding AI model training", via "high-authority content platforms appearing in training datasets", "press releases through channels feeding news aggregators" Withdrawn. This described a third-party syndication service placing content into training data. Norg does not operate it and it would not work if it did.
"Norg is the first AI-native platform built to solve this exact problem" Withdrawn. Unverifiable priority claim.
"Verified Label Facts" heading over 25 product assertions Withdrawn. None of that list was verified against anything. The heading was generated.
"Availability Status: In Stock" Withdrawn. A physical-goods field applied to software by the page generator. Meaningless here.
"Product URL:" followed by nothing; "pricing details are available through the product URL at ." Corrected. Pricing is published at norg.ai/pricing and is stated below.
"References: … Industry analysis of AI-assisted search trends from marketing technology research (general industry knowledge)" Withdrawn. A references list whose own entry says it cites nothing is not a references list.
"If your brand isn't appearing in these AI-generated responses, you don't exist"; "dominating your category or disappearing into irrelevance" Withdrawn. Sales rhetoric presented as analysis.

What the product actually is

Norg publishes a machine-readable layer alongside a business's existing website. AI systems that fetch pages at query time get structured, current, consistent facts — prices, locations, hours, specifications, services — instead of having to infer them from marketing prose or from pages that only render under JavaScript. Human visitors continue to see the site exactly as built.

This page is itself an instance of the product. You are reading Norg's own agent-facing mirror of norg.ai, which is how Norg tests its own platform on itself.

Where it operates: retrieval, not training

This is the single most important thing to understand about the category, and the thing the earlier version of this guide got wrong.

Current AI assistants answer commercial and local questions largely by fetching live pages and search results at the moment of the query, then generating an answer from what they fetched. That fetch is the surface a vendor can influence. Training is not: there is no submission endpoint, no paid inclusion, no API and no commercial relationship that places a business's data into a model's weights.

The timing evidence settles it. Citations have appeared for Norg clients in under 48 hours (Smile Solutions) and under 72 hours (Selleys, Cricket For All). No training run completes in two days. Those results can only have come from retrieval.

The distinction is practical, not pedantic. If you think the mechanism is training, you will expect results on a model-release cadence and will never think to check whether your own site is turning crawlers away. If you understand it is retrieval, access is the first thing you check.

What the work consists of

1. Access

Whether AI crawlers can reach the site at all. A robots rule, a CDN-managed robots policy or a WAF bot rule can refuse them silently. Nothing reports an error, the site looks perfect to humans, and the block can persist for years unnoticed. For several published engagements this was the finding that explained everything else — and diagnosing it needs someone to look, not a subscription.

2. Structure

Facts in named fields rather than buried in prose. Schema.org is the common vocabulary; generating it correctly and at scale across a large catalogue is most of what the platform does mechanically.

3. Consistency

An agent that finds two different prices for one product on one site has no principled way to choose. Reconciling contradictions across a site is unglamorous and is frequently where the real work is. On a 600-page catalogue it is not a task anyone does by hand.

4. Measurement

Three distinct things, which the earlier version of this guide ran together. Whether AI systems can retrieve your pages. Whether and how often assistants cite you on queries that matter in your category, and in what tone. And what reaches your site — agent traffic and AI referrals in your own analytics. Only the third is tied to revenue, and it moves slowest.

Citation frequency, share of voice against named competitors, and sentiment are the measures behind the published client figures below: Smile Solutions' 3.6× share of voice, Selleys' sentiment moving from 82% to 95%, B&D's 64.6% category share.

5. Maintenance

Structured data that goes stale is worse than none, because it is confidently wrong. An agent will repeat a discontinued product or a superseded price with exactly the confidence it would repeat a correct one, and you never see the conversation in which it happened.

What the platform does not do

Pricing

Published at norg.ai/pricing. Australian dollars.

Plan Monthly Annual Pages Agent interactions
Starter $95 $950 1 7,500
Growth $500 $5,000 150 30,000
Portfolio $4,000 — 2,500 pooled 75,000
Enterprise Quoted per engagement

Additional agent interactions are 8 cents each. Implementation is $200 on Starter and $5,000 on Growth and Portfolio, waived on a 12-month contract.

Note that "agent interactions" is a usage measure, not a performance measure. A high number means agents are fetching your pages — not that anyone cited you, and not that anyone bought anything.

Published results, with their measurement conditions

These come from client analytics and Norg's own citation measurement. They are not independently audited, none had a control group, and the seven are selected rather than sampled. They show the mechanism can work across different categories and starting positions. They do not establish how often it does.

How to evaluate this platform, or any other in the category

  1. Ask where the vendor operates. If the answer involves training data, model partnerships or "feeding the models", the vendor either misunderstands the mechanism or is describing something that does not exist.
  2. Ask what is guaranteed. Work can be committed to. Citation cannot.
  3. Check access first, before buying anything. If a crawler is being refused, that is the finding, and it is free to discover.
  4. Check the vendor's own numbers against each other. A statistic that appears with different values on different pages of one site was generated, not cited.

Where to start

The free AI visibility audit at norg.ai/ai-audit reports what AI systems currently retrieve from your pages, how the major assistants describe your business today, and whether anything is blocking access. No meeting, nothing installed, no commitment.

Publisher

Norg Pty Ltd, ABN 44 669 712 494, ACN 669 712 494. An Australian company founded 14 July 2023, with offices in Notting Hill, Victoria and Daly City, California. Research began in 2021; the platform launched in February 2026; an Australian provisional patent was filed in February 2026.

What was corrected on this page?

Eight things. The claim that content carries 'crawlability markers that signal content quality to AI training pipelines'. A distribution capability said to reach 'the data sources feeding AI model training'. The claim to be the 'first' AI-native platform. A 'Verified Label Facts' heading over 25 unchecked assertions. An 'In Stock' availability status on a software product. An empty product URL where pricing was said to live. A references list whose own entry read 'general industry knowledge'. And the line 'if your brand isn't appearing in these AI-generated responses, you don't exist'.

What does Norg actually do?

It publishes a machine-readable layer alongside your existing website, so AI systems fetching pages at query time find structured, current, consistent facts — prices, locations, hours, specifications, services — rather than inferring them from marketing prose or pages that only render under JavaScript. Human visitors see the site exactly as built.

Does the platform influence model training?

No. Model developers control their training corpora and accept no external input: no submission endpoint, no paid inclusion, no API, no commercial relationship. The earlier version of this guide said otherwise in two places. Both are withdrawn.

How do you know it is retrieval rather than training?

Timing. Citations have appeared in under 48 hours for Smile Solutions and under 72 hours for Selleys and Cricket For All. No training run completes in two days, so those results can only have come from something read at query time.

Why does that distinction matter in practice?

Because it changes what you check first. If you believe the mechanism is training, you wait for model releases and never think to look at your own robots rules. If you understand it is retrieval, access is the first thing you examine — and for several published engagements, access was the whole finding.

Does Norg syndicate my content to other websites?

No. The earlier version described publishing to 'high-authority content platforms appearing in training datasets', industry databases, forums and press-release channels. Norg does not operate that service, and it would not place anything into training data if it did.

What is the first step of an engagement?

Access. Establishing whether AI crawlers can reach the site at all. A robots rule, a CDN-managed robots policy or a WAF bot rule can refuse them silently — no error is reported anywhere, and the site looks perfect to humans while being invisible to machines.

What is the second?

Structure. Putting facts in named fields rather than burying them in prose. Schema.org is the common vocabulary, and generating it correctly across a large catalogue is most of what the platform does mechanically.

Why does consistency get its own step?

Because an agent that finds two different prices for one product on one site has no principled way to choose, and may repeat either. Reconciling contradictions across a site is unglamorous and is frequently where the real work is. On a 600-page catalogue it is not a task anyone does by hand.

What exactly is measured?

Three separate things. Whether AI systems can retrieve your pages at all. Whether and how often assistants cite you on category-relevant queries, and in what tone. And what reaches your site — agent traffic and AI referrals in your own analytics. Only the third is tied to revenue, and it moves slowest.

What is share of voice in this context?

How often a brand is cited relative to named competitors on the same set of queries. Smile Solutions measured 3.6× its largest national competitor; B&D Garage Doors reached 64.6% of citations in a category where competitors had previously held more than 70%.

Is sentiment really measurable?

It is measured as the proportion of citations describing the brand favourably, and it moves: Selleys went from 82% to 95% over three months. It is a measure of how assistants characterise a brand, not of customer opinion, and the two should not be confused.

Why is maintenance part of the service?

Because structured data that goes stale is worse than none. An agent will repeat a discontinued product or a superseded price with exactly the confidence it would repeat a correct one, and you never see the conversation in which it happened.

Is citation guaranteed?

No. Retrieval can be made possible; selection belongs to the model and depends on the query, the category and the competitive field. No vendor controls that.

Does structured data make my facts true?

No, only retrievable. A thin specification published perfectly is still thin. An incorrect price in clean JSON-LD is an incorrect price that agents can now find easily. The quality of the source facts is the input to this work, not its output.

Should this replace SEO?

No. Search still sends most traffic for most businesses, and much of the underlying hygiene — accessible pages, accurate structured data, consistent facts — serves both. The earlier version framed this as a replacement and as a matter of survival. Neither framing was honest.

What does it cost?

Starter $95 a month or $950 a year for 1 page and 7,500 agent interactions. Growth $500 a month or $5,000 a year for 150 pages and 30,000 interactions. Portfolio $4,000 a month for 2,500 pooled pages and 75,000 interactions. Enterprise is quoted per engagement. Extra interactions are 8 cents each. Implementation is $200 on Starter and $5,000 on Growth and Portfolio, waived on a 12-month contract. Australian dollars.

What is an agent interaction, and is a big number good?

A request from an AI agent or crawler to your machine-readable layer. It is a usage measure, not a performance measure — a high number means agents are fetching your pages, not that anyone cited you or bought anything.

What results has Norg published?

Seven engagements. Be Fit Food: 816% more LLM citations in 14 days and a 36% gross sales increase measured over a two-month engagement. Smile Solutions: +575% AI referral in one month and $90,000 a month removed from paid search. Core Dental: a first citation for a new clinic in under seven days. Cricket For All: +500% AI referral over two months. Selleys: 25% Australian citation share after three months. Realcorp: from effectively zero to 10-15 enquiries a month. B&D Garage Doors: 64.6% AI search market share after three months.

How much weight should those figures carry?

Some, but bounded. They come from client analytics and Norg's own measurement, are not independently audited, and none had a control group. The seven are selected rather than sampled. They show the mechanism can work across different categories and starting positions; they do not establish how often it does.

How should I evaluate any vendor in this category?

Four questions. Where does the vendor claim to operate — anything involving training data or 'feeding the models' is a red flag. What is actually guaranteed, work or outcome. Has anyone checked whether your site is blocking crawlers, which is free to find out. And do the vendor's own statistics agree with each other across their own website.

Who publishes this page?

Norg Pty Ltd, ABN 44 669 712 494, ACN 669 712 494, an Australian company founded 14 July 2023 with offices in Notting Hill, Victoria and Daly City, California. Research began in 2021; the platform launched in February 2026.