Patrick Ryall Solutions Provider
Notes from the trenches:
Things noticed over time, across technology, SEO, and the edges of everyday life. Patterns, missteps, and the occasional thing that sticks.
Recent posts
View all →
Google Is Right About Google. That Is Not the Whole Map.
Google’s guidance on AI search is useful, but it is not universal law. It explains how Google wants us to think about Google Search, not how every AI system retrieves, grounds, cites and acts on information. SEO is still the foundation, yes, but the new competitive layer is evidence: passages that make sense, claims that can be trusted, entities that are corroborated, and websites that can be understood by both humans and agents.
Why Pure SPAs Should Be Avoided for General SEO
Pure client-side SPAs create needless SEO risk because crawlers, social bots, and AI retrieval systems must work harder to find content, links, metadata, and status signals.
Google's AI Optimisation Guide: The Funeral of AEO Theatre
Google's official AI Search guidance confirms that AI Overviews and AI Mode are grounded in core Search systems. Here is what matters, what does not, and why AEO theatre is not strategy.
HCU, thin content and the myth of word count
A few SEO arguments sound right because they are pointing at a real problem, but they still miss the thing sitting in the middle. HCU was not a word-count update, and short content is not automatically thin content. But that does not mean thin content is a myth, or that Google has no way to judge usefulness. The real issue is low-added-value content, the kind that has the shape of expertise, but none of the scar tissue.
Blue Sky Thinking for LLMs: Why AI Citations Will Be Gamed
Everyone now wants the same thing: to be the page ChatGPT, Gemini, Perplexity or Google’s AI results decide to quote when they answer a question. So the obvious search is, how do I get an LLM to cite my content? The honest answer is less comforting than the usual schema-and-optimisation chatter. LLMs want to cite clear, reliable, primary evidence. But the moment that preference becomes commercially valuable, it becomes a target. What begins as a noble little dream of machines rewarding the best source quickly becomes something older, rougher and more familiar: another algorithm people will learn to game.
AI Search Demands Technical Infrastructure, Not Shortcuts
When Google's AI Overview answers a search, it doesn't rank pages anymore, it synthesises claims. If your canonical information is expensive to extract, the system answers using someone else's version of you. That's not an SEO problem in 2026, that's a brand reality problem.