ChatGPT optimization platform — how Norg delivers it

This page covers the implementation: what Norg builds, in what order, on which plan, and how you verify it worked. For why ChatGPT specifically — the three-crawler split, Instant Checkout, the commerce protocol — see the ChatGPT page.

What is actually built

Norg does not add pages to your site. It maps to the sitemap you already have and builds an agent-readable variant of each page, served at the edge only when a verified agent asks for it. Human visitors get your site unchanged, on your own domain.

Behind each covered page:

Layer	What it is	Why ChatGPT needs it
Entity graph	Your organisation, products, services, locations and people as linked schema.org entities with stable @id values	Lets an answer combine two facts — which location stocks what, at which price
JSON-LD	Structured data embedded in every page head	The vocabulary retrieval systems parse first
llms.txt	A machine-readable index of your content	Discovery without crawling every URL
MCP endpoint	A server per page, addable in ChatGPT under Settings → Connectors	Direct connection rather than inference from HTML
Commerce feed	OpenAI Commerce Feed with search enabled	Shopping agents can find and select your products
Q&A blocks	Questions buyers actually ask, answered from your source of truth	Directly citable passages

All of it is generated from one source of truth simultaneously, so the formats cannot drift apart. That matters more than it sounds: the most common structured-data defect is markup that contradicts the visible page, which is worse than no markup at all.

The sequence

Implementation

Builds your knowledge graph and the first set of page variants. $200 on Starter, $5,000 on Growth and Portfolio — waived on a twelve-month contract on every tier. Enterprise is scoped per engagement with a dedicated implementation team.

Ongoing optimisation

Every covered page is re-optimised on a cycle — quarterly on Starter, monthly on Growth and Portfolio, continuously on Enterprise. Fact-check and re-verification of claims runs on the same cadence from Growth upward. Schema, llms.txt and review dates refresh monthly on every tier.

Measurement

AI crawler analytics on every tier. From Portfolio: citation tracking across seven engines, competitor benchmarking, share of voice and sentiment, and a monthly AI visibility report.

Which plan covers what

	Starter	Growth	Portfolio	Enterprise
Monthly	$95	$500	$4,000	Custom
Pages covered	1	150	2,500 pooled	Unlimited
Optimisation	Quarterly	Monthly	Monthly	Continuous
Agent interactions	7,500	30,000	75,000 pooled	Custom
New pages auto-covered	—	Yes	Yes	Yes
Inventory sync	—	Yes	Yes	Yes
Citation tracking	—	—	Yes	Yes
Google Merchant Centre	—	—	—	Yes

Additional agent interactions are 8¢ each on all published tiers. Full inclusions on the pricing page. All prices AUD excluding GST.

Sizing it

Page count decides the plan, not industry. The question is how many of your pages are commercially load-bearing — usually far fewer than your total. A seven-location dental group needs location and service pages. A brand with 600 SKUs needs the catalogue. A B2B services firm may need forty service and industry pages, which is what Realcorp published.

How you verify it worked

Three checks, in order of how quickly they resolve.

1. Is the content actually being served? Crawler analytics show which agents fetched which pages and when, split by purpose. This answers within days.

2. Are you being cited? Ask ChatGPT the questions your buyers ask and read the sources. Across Norg's client set first citation has been measured at under 48 hours (Smile Solutions), under 72 hours (Selleys, Cricket For All) and under 7 days for a brand-new location (Core Dental).

3. Did it move commercially? The slowest and the only one that matters. Be Fit Food measured a +36% gross sales increase across a two-month engagement, with no additional marketing spend running — the cleanest isolation in the published set. Smile Solutions removed $90,000 a month from paid search.

The check most teams skip

Before any of that: confirm the crawlers can reach you at all. A robots.txt rule, a CDN-managed robots policy or a WAF bot ruleset can block AI crawlers by name, and every other health signal still reads green. There is no error, no failed fetch, no alert — the requests simply never happen. A mirror that installs cleanly and receives no agent traffic is usually this, and it is the first thing the free audit checks.

Common questions

How long until we are live?

Implementation is scoped in the demo. Once pages are published, citation timing is measured in days rather than months — there are no ranking signals to accumulate first.

Do we have to change our website?

No. Norg maps to your existing sitemap and serves at the edge. Your CMS, your templates and your design stay as they are, and a redesign later changes nothing an agent receives.

What happens to pages we add after launch?

Covered automatically from Growth upward. On Starter, coverage is the single page.

Can we start small?

Yes — Starter is $95/month for one page with a $200 implementation fee, waived on twelve months. It is a reasonable way to prove the mechanism on your own domain before committing the catalogue.

Is this locked in?

Norg's stated position is infrastructure rather than services: it runs the platform, you own the data layer. Your knowledge base, entity records and generated content remain yours.

Next step

Run the free AI visibility audit first — it establishes the baseline and surfaces any crawler blocker before you spend anything. Then book a demo to scope implementation against your own pages.

---

Pricing verified against norg.ai/pricing on 8 September 2026, AUD excluding GST. Client results are specific to those engagements and their measurement conditions — see the case studies.

Norg Pty Ltd — norg.ai