"We'll publish one AI article a day for you."
If you run a business, you've probably been offered some version of that. More pages, more keywords, more chances to show up. It sounds like momentum. For most businesses, it's the fastest way to bury the pages that actually matter.
What Google actually says
Google's spam policies include a section called scaled content abuse. The definition is one sentence:
"Scaled content abuse is when many pages are generated for the primary purpose of manipulating search rankings and not helping users."
Read it again and notice what it doesn't say. It doesn't ban AI content. It doesn't set a limit on how many pages you can publish. Google is clear the policy applies whether the pages were produced by a person, a tool, or both.
The test isn't who wrote the page. It's why the page exists. A human writer can fail it. A model can pass it.
The bigger problem isn't Google
Most of the conversation about content volume is about rankings. That matters, but it's no longer the whole picture.
When someone asks ChatGPT, Gemini or Claude to recommend a business, the answer doesn't come from page ten of the search results. AI assistants lean on a small set of sources they already trust, and they cite those sources because they know something the model can't find anywhere else.
That changes the math. Three hundred posts don't give you three hundred chances to be cited. If most of them say what everyone else has already said, they give the two or three pages that could have earned a citation a smaller share of your own authority.
What this looks like in practice
We recently went through the blog of a well-run software company. A few hundred posts, written by real people, with named authors and proper sources. By most standards, a content program done right.
Here's what we found:
- Two articles competing for the same question. One was a 3,000 word post answering it. The other was a longer post that asked the same question as one of its headings, then asked it again in its FAQ.
- The numbers doing the heavy lifting were the same industry statistics that appear on thousands of other blogs. In the posts we read, not one figure came from the company itself.
- A glossary of around a hundred short pages, most under 300 words, most built on the same template.
- Buried in the middle of it all, the company's own annual research report. Original data nobody else had. Their site linked to it once. A post restating a well-known survey was linked from almost thirty pages.
Nothing in that library was spam. It was well-intentioned volume. But the one asset an AI assistant would actually want to cite was the hardest thing on the site to find.
A few honest caveats
Google doesn't publish a page count that triggers the policy, so anyone quoting a number is guessing. The traffic losses being reported after the March 2026 core update are third-party estimates, not figures Google has confirmed. What's certain is the policy itself, and the direction AI search is moving.
What to do instead
Simplicity and restraint win. Before publishing anything, ask whether it tells someone something they can't get elsewhere: your pricing, your process, your results, your data, your experience.
- 1Find the pages that compete with each other and merge them into one stronger page.
- 2Put your original work where people and AI can find it, and link to it from the pages that get traffic.
- 3Stop publishing pages that only restate what's already out there.
- 4Check which of your pages AI assistants actually cite, and build on those.
One page carrying data only you have will do more than thirty pages summarising what everyone already published.
Publish less. Know more.
Want to see which of your pages AI assistants cite today? Run your first scan free.