Sydney Business Web

29/08/2026

Have you ever given ChatGPT, Gemini or Grok a link to your website—only to be told that it cannot read the page?

The obvious conclusion is that your website must be blocking AI. But that conclusion may be completely wrong.

We experienced this on our own Cloudflare-protected website. A live AI assistant failed to retrieve the page, yet our AI Observatory showed recognised search and AI crawlers successfully accessing and retrieving business information from the site.

At the time of the test, Cloudflare recorded an anonymous AWS-hosted request targeting the exact article. It used the generic user-agent `axios/1.19.0`, provided no verifiable AI identity and was blocked before reaching WordPress.

That did not mean AI crawlers could not crawl our website. It meant one live retrieval request failed.

It also did not mean we should turn off Cloudflare or weaken our security. The correct response is to inspect the evidence, establish what was blocked and make only the narrowest change that can be technically justified.

Our new Technical Problem 16 explains what happened, what the evidence proves and what website owners should do when an AI says it cannot read their website:

https://sydneybusinessweb.com.au/why-ai-cannot-read-your-website/

28/08/2026

Old Father SEO isn’t being packed off to a retirement village. AI Visibility is simply taking him by the hand and walking him into the future.

Traditional SEO remains essential—but search is no longer limited to ranking blue links. AI-assisted systems now retrieve information, connect evidence, compare businesses and sometimes make recommendations inside the answer itself.

Google says there is no special AI markup or magic formula. True. But that does not mean every website gives machines an equally clear account of who the business is, what it does and why it should be trusted.

Our new article explains how SEO and AI Visibility work together, why coherent business entities matter to ordinary Google Search too, and where specialist tools reveal problems conventional reports cannot see.

It contains engineering, evidence, Old Father SEO—and absolutely no ceremonial invoices. 😄

https://sydneybusinessweb.com.au/ai-visibility-and-seo-search-has-changed/

22/08/2026

We’ve just published a new real-world Schema Gorilla case study using the Sydney Business Web website itself.

The fresh analysis examined 273 pages, reconstructing 14,577 schema nodes and 36,414 relationships across the whole site.

The interesting part is that this was not our first run.

An earlier analysis found structural and entity issues. We corrected them, continued expanding the site and its machine-readable identity, then ran Schema Gorilla again from scratch.

The new result: zero findings under ruleset 2.2.

The article explains why checking individual pages for valid schema is only part of the problem. A business can have individually valid pages while still presenting fragmented or conflicting identities when the whole website is viewed as one connected system.

Full case study:

Schema Gorilla Case Study: a fresh analysis of 273 pages found 14,577 schema nodes, 36,414 relationships, 28 entity clusters and zero findings.

20/08/2026

Are businesses still spending money on the wrong search problem?

SEO still matters. Rankings still matter. Useful content still matters.

But AI systems can now retrieve business information, compare suppliers and generate answers without necessarily sending the user through to a website.

So there are new questions businesses need to ask:

Can AI systems actually retrieve our information?
Do they encounter one coherent business identity?
Can they connect us with an unbranded customer need?

Our new article uses real Google AI evidence together with results from AI Observatory and Schema Gorilla to look at the difference between AI Visibility and traditional SEO.

And yes — we deliberately challenge the economics of continuing to spend heavily on yesterday’s measurement model without checking what today’s search systems can actually see.

https://sydneybusinessweb.com.au/ai-visibility-vs-seo-wasting-money/

15/08/2026

We’ve now published the full page for Schema Gorilla — Business Identity Analysis for AI. 🦍

Schema Gorilla was built to solve a problem ordinary schema validators can easily miss:

individual schema blocks can be valid while the website as a whole still describes the business inconsistently.

Instead of examining isolated pieces, Schema Gorilla connects structured data from across the website and looks at the complete machine-readable business — the organisation, people, services, content, identities and relationships between them.

Our production commissioning run on Sydney Business Web analysed 278 pages, 13,569 schema nodes and 33,896 relationships.

It found five underlying structural problems. We corrected them, ran a completely fresh crawl and rebuilt the graph.

Final result: zero detected findings.

Have a look here:

https://sydneybusinessweb.com.au/schema-gorilla-business-identity-analysis-for-ai/

12/08/2026

We’ve just published a detailed look at our new AI Observatory — Verified AI Retrieval Monitoring.

There is a huge difference between saying “an AI bot visited the website” and demonstrating that an identifiable AI system successfully retrieved useful business content.

Our monitoring filters out much of the surrounding machine noise and focuses on the retrieval activity that actually matters.

It’s the first measurable layer in the bigger AI visibility chain:

Retrieval → Understanding → Mention/Recommendation → Citation

Have a look here:

https://sydneybusinessweb.com.au/ai-observatory-verified-ai-retrieval-monitoring/

11/08/2026

AI Visibility is mostly invisible engineering.

A lot of the work happens under the hood — entity architecture, structured data, crawler access, evidence structures, retrieval pathways and filtering.

That can make AI Visibility feel a bit abstract.

So we’ve built something that makes the first measurable layer tangible.

The new Sydney Business Web AI Observatory shows:

Who came?
What did they retrieve?
Did the retrieval succeed?

This is Level 1: retrieval.

It doesn’t prove that an AI system has understood, recommended or cited your business. But it gives you a genuine peek under the hood at the first observable stage of the process.

And importantly, we filter out a lot of meaningless bot noise rather than simply counting every automated visit.

We’re now preparing the Observatory for external business trials.

https://sydneybusinessweb.com.au/ai-observatory/

10/08/2026

Doncha love it when stuff works?

This data is generated from edge Workers running on Cloudflare. The requests are observed at Cloudflare’s edge before they reach the website origin, and the resulting retrieval outcomes are recorded there too.

The telemetry is then filtered, corroborated and categorised so we get useful numbers for the actual business pages and machine-discovery resources that credible AI and search crawlers are retrieving — not useless raw “Kilroy was here 6,000 times” bot-hit numbers.

Synthetic tests, diagnostic fetches, support assets and weakly identified crawler traffic are excluded from the headline evidence.

Even the telemetry stays in Cloudflare: it’s stored in Analytics Engine and queried using SQL.

The point isn’t to count bots. It’s to measure what identifiable machine systems actually retrieve — and whether the retrieval succeeds.

10/08/2026

We’ve now put our AI retrieval measurement system live.

Rather than simply saying a website is “AI-ready” or that certain crawlers are allowed, we wanted to measure what actually happens at the website.

Which search and AI-related machines reach it?

What do they retrieve?

Does retrieval succeed?

And can we distinguish credible machine activity from tests, support assets and crawler noise?

The result is our new Live AI Retrieval Evidence page.

The figures update against a rolling 24-hour observation window, so the systems shown can change as real machine activity changes.

👉 https://sydneybusinessweb.com.au/ai-retrieval-evidence/

This is not intended to prove that an AI understood, cited or recommended the business. It measures something earlier in that chain:

observable machine access and retrieval.