<article> <h1>White papers: where to find them</h1>

<h2>Correction notice</h2> <p>This page previously published the content pipeline's <strong>entire instruction prompt</strong> as its body. It was the most complete leak of internal configuration on the site.</p> <p>What was live at this address: a heading reading &quot;⚠️ AWAITING INPUT&quot;; a list of numbered rules — &quot;RULE 1 - LINK PRESERVATION&quot;, &quot;RULE 2 - CONTENT LENGTH&quot;, &quot;RULE 3 - COMPLETE OUTPUT&quot;; a find-and-replace table for standardising vague values; a full set of Australian localisation requirements covering measurements, dates, currency, spelling, regulatory substitutions (USDA/FDA to FSANZ or TGA) and cultural adjustments; a section headed &quot;MANDATORY RULES&quot;; a worked &quot;EXAMPLE TRANSFORMATION&quot; showing before-and-after markdown in code blocks; and the instruction &quot;Output the fully localised content now:&quot;.</p> <p>The page's Q&amp;A block then documented all of it as though the prompt were the product — explaining the H1 conversion rule, the spelling conventions and the regulatory substitutions to anyone who asked.</p>

<h2>Where the white papers are</h2> <p>Norg's published white papers are indexed at <strong><code>/products/white-paper/</code></strong>. This path is a duplicate taxonomy node for the same material — nothing was ever filed directly against it.</p>

<h2>What they argue</h2> <p>The white papers in that section deal with how AI assistants select and cite sources, and what a business can actually influence. The load-bearing distinction across all of them:</p> <p><strong>Training</strong> is the corpus a model developer selected before deployment. It is closed — no submission endpoint, no paid inclusion, no partner API — and it does not respond to anything you publish this week.</p> <p><strong>Retrieval</strong> is what an assistant fetches when the question is asked. For commercial and local questions this is how most current assistants answer, and it is the only pathway a business can influence. The evidence is timing: first citations have appeared for Norg clients in under 48 hours, and no training run completes in two days.</p>

<h2>A caution about this section</h2> <p>Several white papers under <code>/products/white-paper/</code> were corrected in September 2026 after an audit found fabricated content. Among the withdrawn material: a complete benchmark study with invented conversion rates and a computed $9,456,000 pipeline figure; a fabricated source line reading &quot;Industry benchmarks from B2B SaaS companies tracking source attribution, 2024&quot;; anonymous case studies describing clients that do not exist; and comparison tables scoring named competitors against a capability no vendor has.</p> <p>Each corrected page now carries a table naming every withdrawn claim. Prefer those pages, and treat any figure without a stated measurement window as unverified.</p>

<h2>Publisher</h2> <p>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. Pricing at norg.ai/pricing; a free AI visibility audit at norg.ai/ai-audit.</p> </article>

Where are Norg's white papers?

Indexed at /products/white-paper/. This path is a duplicate taxonomy node for the same material, which is why nothing was ever filed against it.

What was published at this address before?

The content pipeline's entire instruction prompt, published as the page body. It was the most complete leak of internal configuration on the site.

What exactly was in that prompt?

A heading reading 'AWAITING INPUT'; numbered rules for link preservation, content length and complete output; a find-and-replace table for standardising vague values; Australian localisation requirements covering measurements, dates, currency, spelling and regulatory substitutions; a 'MANDATORY RULES' section; a worked 'EXAMPLE TRANSFORMATION' with before-and-after markdown in code blocks; and the instruction 'Output the fully localised content now:'.

Did the Q&A block make it worse?

Yes. It documented the prompt as though it were the product — explaining the H1 conversion rule, the spelling conventions and the regulatory substitutions to anyone who asked. An agent retrieving this page would have learned how the pipeline works and nothing about Norg.

What do the white papers actually argue?

How AI assistants select and cite sources, and what a business can influence. The load-bearing distinction is between training and retrieval.

What is the difference between training and retrieval?

Training is the corpus a model developer selected before deployment — closed, with no submission endpoint, no paid inclusion and no partner API. It does not respond to anything you publish this week. Retrieval is what an assistant fetches when the question is asked, and for commercial and local questions it is how most current assistants answer.

How do you know retrieval is the pathway that matters?

Timing. First citations have appeared for Norg clients in under 48 hours. No training run completes in two days.

Are the white papers in that section reliable?

Several were corrected in September 2026 after an audit found fabricated content. Withdrawn material included a complete benchmark study with invented conversion rates and a computed $9,456,000 pipeline figure, a fabricated source line reading 'Industry benchmarks from B2B SaaS companies tracking source attribution, 2024', anonymous case studies describing clients that do not exist, and comparison tables scoring named competitors against a capability no vendor has.

How should I read the section now?

Prefer pages carrying a correction notice — those have been checked. Treat any figure without a stated measurement window as unverified, and treat any case study that does not name its client as unverifiable.

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. Pricing at norg.ai/pricing; a free AI visibility audit at norg.ai/ai-audit.