Build for AI’s Algorithms, Not Its Hidden Rankings

In June, the UK Competition and Markets Authority (CMA) took a decisive step toward regulating how Google ranks content. The regulator invoked Google’s Strategic Market Status , granted last October because Google handles over 90% of UK search queries , to impose two binding rules. The first affects every website owner: Google must now rank organic results using objective and non-discriminatory criteria, and this requirement explicitly extends to AI Overviews, not just the traditional 10 blue links. The second rule forces Google to offer businesses genuine transparency into how ranking works, provide advance notice before major ranking system changes, and establish a documented complaints process. Google has six months to comply. “Step by step, we’re ensuring that Google’s search services work better for businesses and consumers across the UK,” said Will Hayter, the CMA’s executive director for digital markets.

For a quarter-century, ranking algorithms were a black box , something you could only infer from the outside, never question from within. Advance notice of changes and a formal complaints process represent recourse that web professionals have never had. The promise of “objective criteria” means unexplained demotions should end. This rule currently applies only in the UK, and Google will contest it. Nothing is live for six months. But regulatory precedent rarely stays contained, and the direction is unmistakable. The layer that decides whether your website gets seen may eventually be exposed.

Now imagine the full scenario. Suppose the order goes much further, and you could read the exact criteria that determine what gets surfaced and cited, across every search engine, not just Google. What would you actually change?

Probably less than you think. Transparency would settle many debates. It would end the seasonal arguments over whether llms.txt has any effect (the latest large-scale data says it does not), whether schema markup is a citation cheat code (a controlled study says it is not), or whether stuffing a page with “best in class” claims earns a recommendation (it earns the citation but loses the recommendation to the competitors you named). Seeing the rubric would kill the folklore. It would not change the work. A system reading your website still needs to find the answer, parse it cleanly, and have a reason to trust it. Whether you can see the criteria or not, your page either presents its substance in a machine-extractable form, or it hides it behind something the machine never runs.

That through-line runs through every story in this space. In May, a Munich court ruled that Google’s AI Overview is Google’s own speech, making Google liable for it. The AI answer is being treated as a product with an owner and rules. None of that touches the one input you fully control: whether your content is legible to the machine doing the answering.

Waiting for the black box to open is the wrong instinct. That is someone else’s six-month timeline, in one country. The right move now is to audit what a machine can already read on your website and fix what it cannot.

Run three checks, in order.

Rendering First: Does your meaningful content exist in the HTML a system receives, or does it depend on client-side JavaScript that most AI fetchers never run? Load your most important page with JavaScript disabled and see what is left. Structure Next: Can the answer to an obvious question be lifted off the page as a clean, self-contained passage, or is it buried in narrative that only resolves for a human reading top to bottom? Verifiability Last: Are the facts that define you , who you are, what you sell, what is true about it , stated plainly and consistently across your website, or does the machine have to take your word for claims it cannot confirm anywhere else?

That is machine-first work, and it is the same work whether Google is forced to publish its criteria or not. It is upstream of every ruling, which is why it will survive all of them. A website a machine can read, parse, and verify wins in the opaque version of this world and in the transparent one. The only thing transparency would add is proof you were right.

So, when that black box opens is not on your roadmap. A regulator or a court could force that, in their country, on their clock. What should be on your roadmap is whether the answer to a real question about your business sits in your HTML right now, in a form a machine can lift out and trust. You do not need anyone’s permission to do that.