If your agency is charging you for blog content, there is a decent chance they are generating it with AI, running it through a "humanizer," and calling it a deliverable. And for the last couple of years, the industry argument about that has been the wrong one: can Google detect AI content or not?
Ann Smarty just published a sharp piece on this, and her framing matches what I have been telling moving company owners for a while. Google does not need to detect AI. It is scoring something much simpler and much harder to fake: effort. And there is real documentation behind this, not just SEO Twitter speculation. Let me walk you through the receipts, because this changes how you should think about every content deliverable you pay for.
The word "effort" appears over 100 times in Google's own guidelines
Google publishes a document called the Search Quality Rater Guidelines, the handbook given to the thousands of human raters who evaluate whether Google's search results are any good. The word "effort" shows up in it more than a hundred times. Here is the definition, verbatim:
Effort: The extent to which a human being actively worked to create satisfying content.
The guidelines get specific about what does not count. Running other people's content through a translator or a generator and publishing thousands of pages with no oversight or curation is their example of no effort. And here is the part that should worry anyone buying cheap content: since the January 2025 revision, which Search Engine Land covered at the time, the guidelines say content with little to no effort, little to no originality, and little to no added value should be rated Lowest. Citing your sources does not rescue it. Attribution is not effort.
One more quote from the same document, because it cuts both ways and I want to be honest about that:
The use of Generative AI tools alone does not determine the level of effort or Page Quality rating.
Google is not anti-AI. Google is anti-lazy. Those are different wars, and only one of them is coming for your agency's blog package.
The leaked documents put a name on it: contentEffort
In May 2024, thousands of pages of Google's internal Content Warehouse API documentation leaked, and SEOs have been mining it ever since. Buried in there is an attribute literally named contentEffort, described in the documentation as an "LLM-based effort estimation for article pages."
Read that again. Google uses a language model to estimate how much human effort went into your article. The thing agencies are using to avoid effort is being used by Google to measure it.
Now the honest caveat, because I am not going to sell you hype: the leak proves the attribute exists. It does not prove how heavily it is weighted in live rankings, and nobody outside Google can tell you that number. What it proves is direction. Google built machinery to score effort at web scale, and the rater guidelines tell you exactly what that machinery is trained to look for.
"We reward high-quality content however it is produced" was never a green light
When AI writing tools took off, Google published guidance saying it rewards high-quality content, however it is produced. A lot of agencies read that as permission. It was not. It was a warning with good PR.
The same company's spam policies define scaled content abuse as producing many pages "for the primary purpose of manipulating search rankings and not helping users," and the policy applies no matter how the pages were created. Put those two statements together and the actual position is obvious: they do not care what tool you used. They care whether a human being added something real.
That is why the "undetectable AI content" pitch is snake oil even when it works. Passing a detector was never the test. The test is whether anything on the page could not have been produced by paraphrasing what already ranks.
What this looks like for a moving company
Here is the question I want you to ask about any content on your site, whether your agency wrote it, your office manager wrote it, or you wrote it yourself: what is on this page that only someone who knows this business could have written?
A real answer sounds like this. The actual deposit policy you enforce and why. What a two-bedroom apartment move really costs in your city this season, not a national average scraped from a blog. The building downtown that requires a certificate of insurance before your crew can touch the elevator, and what happens when a mover shows up without one. How you handle a claim when something breaks. The reason your minimum hours exist.
None of that is in the top ten results for any keyword. That is exactly the point. It cannot be generated from the results page, so it cannot be replicated by someone with a prompt. That is effort in Google's vocabulary, and it is the same standard I apply to every post on this site, including the breakdown of content mistakes that tank moving company rankings and how AI search actually works for movers.
Effort is not what you think it is
It is not word count. A 3,000-word article that paraphrases ten ranking articles is low effort at any length. It is not research depth either, because AI does research fine. It is not the time you spent refining your prompt. Nobody can see your prompt, and Google does not grade your process.
Effort is what survives onto the finished page that a stranger can point at. A specific number from your market. A named example someone could verify. A position stated plainly enough to disagree with. A photo from your actual job site instead of a stock image. One of those is worth more than a thousand extra words.
So what do you actually do about it
If you are paying an agency for content, audit it with the question above. If the articles could run on any moving company site in any city with the name swapped, you are buying Lowest-rated content, and the question of whether your agency is worth the money has a pretty clear answer.
If you write your own content or work with someone who does, the playbook is boring and it works. Fewer posts. Real details from your real operation. Your pricing logic, your service area quirks, your worst job and what you learned from it. AI can help you outline, organize, and edit. It cannot supply the parts that matter, because the parts that matter are locked inside your business.
That is the whole opportunity here, honestly. When every competitor can generate fifty articles a month, the bar is not volume. The bar is knowing something the generator cannot know. As a working mover, you already have that. The only question is whether it makes it onto the page.
If you want a content strategy built around what only you know, that is literally what we do. No content farms, no humanizer tricks, no fifty-post packages. Just your expertise, on the page, where Google is actually looking for it.