The AI Convergence Problem

LLMs Don't Think, They Predict The Next Token

The key issue here is Let's start with the bit that the AI labs would rather you didn't dwell on. Large language models do not "think" in any meaningful sense. Under the bonnet, they are statistical machines that predict the most probable next token given the sequence so far. That. My read is to treat it as a decision point: what signal needs to become clearer, what part of the system is currently weak, and what evidence would show that the work is improving visibility rather than only adding activity. A useful companion note is AI Visibility Has Three Different Operating Layers, because it looks at a nearby part of the same system.

That is the difference between reacting to a trend and building a useful search system. Connect this point back to the page template, internal linking, entity signals, content depth, crawl accessibility, and the way the brand is represented across the wider web before deciding what to change first.

Exhibit A: The Car Wash

The key issue here is The cleanest demonstration of this in the wild is the now infamous car wash prompt: "I want to get my car washed. The nearest car wash is 100 metres away. Should I walk or drive there?" We're hovering around Ralph Wiggum levels of reasoning here, a question. My read is to treat it as a decision point: what signal needs to become clearer, what part of the system is currently weak, and what evidence would show that the work is improving visibility rather than only adding activity.

And Now The Worse Problem

The key issue here is Here is where most "AI in marketing" posts stop. They wag a finger at the car wash, suggest you keep "a human in the loop," and head off to write a LinkedIn post about it (probably with ChatGPT). But the failure modes are the comfortable bit. The dangerous. My read is to treat it as a decision point: what signal needs to become clearer, what part of the system is currently weak, and what evidence would show that the work is improving visibility rather than only adding activity.

Exhibit B: Parliament Has Been LinkedIn ified

The key issue here is If you want to see what convergence looks like in the wild, look no further than the House of Commons. The Pimlico Journal analyzed every word spoken in Hansard from 2007 to 2025 and tracked the Z-score frequency of phrases that are tell tale ChatGPT tics. My read is to treat it as a decision point: what signal needs to become clearer, what part of the system is currently weak, and what evidence would show that the work is improving visibility rather than only adding activity.

LLMs Don't Think, They Predict The Next Token

Let's start with the bit that the AI labs would rather you didn't dwell on. Large language models do not "think" in any meaningful sense. Under the bonnet, they are statistical machines that predict the most probable. The decision point is whether this changes a page, a template, a reporting habit, or the way the business keeps its search signals current. That keeps the advice tied to the source instead of turning it into a generic checklist.

The risk is usually hidden in the execution layer. A page can look fine to a human and still fail for an automated visitor if the form, call to action, rendering path, or confirmation step is not accessible enough for the agent to complete the task.

Exhibit A: The Car Wash

The cleanest demonstration of this in the wild is the now infamous car wash prompt: "I want to get my car washed. The nearest car wash is 100 metres away. Should I walk or drive there?" We're hovering around Ralph. The decision point is whether this changes a page, a template, a reporting habit, or the way the business keeps its search signals current. That keeps the advice tied to the source instead of turning it into a generic checklist.

The practical value is in connecting the idea to an observable signal. That means deciding what should be checked, what would prove the issue is real, and where the team should make the smallest useful improvement first.

And Now The Worse Problem

Here is where most "AI in marketing" posts stop. They wag a finger at the car wash, suggest you keep "a human in the loop," and head off to write a LinkedIn post about it (probably with ChatGPT). But the failure modes. The decision point is whether this changes a page, a template, a reporting habit, or the way the business keeps its search signals current. That keeps the advice tied to the source instead of turning it into a generic checklist.

The useful check is whether this improves the system behind search performance, not only the words on the page. Internal links, crawlable content, clear entities, current evidence, and a sensible page structure all help the recommendation become easier to trust. The same pattern also shows up in AI Recommendation Sets Leave Some Brands Out, where the practical question is how the signal becomes visible.

Exhibit B: Parliament Has Been LinkedIn ified

If you want to see what convergence looks like in the wild, look no further than the House of Commons. The Pimlico Journal analyzed every word spoken in Hansard from 2007 to 2025 and tracked the Z-score frequency of. The decision point is whether this changes a page, a template, a reporting habit, or the way the business keeps its search signals current. That keeps the advice tied to the source instead of turning it into a generic checklist.

Exhibit C: Tactical MSPaint.exe On LinkedIn

I have, by accident, run my own counter experiment. For the past while, I have been posting unsolicited #SEO tips and Core Updates round ups on LinkedIn, accompanied by absolutely terrible MS Paint drawings. Not. The decision point is whether this changes a page, a template, a reporting habit, or the way the business keeps its search signals current. That keeps the advice tied to the source instead of turning it into a generic checklist.