A few weeks ago, Google quietly made an important change to how they rank content: they split content into two categories: commodity, and non-commodity content.
Commodity content is everything an AI Overview can summarize away in three sentences.
Non-commodity content is the stuff people still bother to click on.
The line was already there. They just named it.
This matters more for AI marketing content than for any other kind, because AI is exactly what made commodity content cheap to produce at scale.
The race to publish “10 ChatGPT prompts” articles has produced thousands of pages that read the same, rank for nothing, and earn zero clicks. (I’ve ‘written a few of them. We all have.)
This morning I was reading Harry Clarkson-Bennett’s article for Search Engine Journal. The more I thought about it, the more I realized this is the single biggest framing change for anyone writing AI marketing content in 2026.
This article is about what AI content that ranks actually looks like now, why most AI content is already invisible.
What does it mean for AI marketing content to be “invisible”?
Invisible AI marketing content is content that LLMs can summarize so cleanly that a human never needs to click through to the source. The answer is delivered. The site is uncredited. The traffic is gone.
This is what zero-click search produces at scale. And it is what the entire SEO industry is currently rebuilding around.
The mechanism is simple. If your post is “10 ChatGPT prompts for marketing” and the prompts themselves are pulled from public Reddit threads, an AI Overview can stitch them together with citations to higher-authority sites. Your version of the same answer adds nothing. Google’s information gain patent (US20200349181A1) literally scores the added value a page contributes compared to other documents on the same topic. If your score is low, your page becomes raw material for the summary, not the destination after it.
I wrote about a version of this in The AI Listicle Loophole, where “best of” pages built for affiliate revenue are getting eaten alive in AI search results. The same dynamic applies to AI marketing content broadly. If your AI content is a remix of what already exists, AI tools will treat it as remixable. That is the loop.
Why is commodity AI content suddenly dying?
Commodity AI content is dying because it does two jobs poorly at once. It can be summarized perfectly by an answer engine, and it cannot make money in a world where users never click through.
Google’s own developer documentation now reads, almost word for word: “Focus on making unique, non-commodity content that visitors from Search and your own readers will find helpful and satisfying.”
That is the polite version. The blunt version is that commodity content was already on borrowed time, and AI tools made it so easy to produce that the floor collapsed under it.
Here are the two questions Harry suggests anyone in marketing should ask before publishing a piece. I have adapted them slightly for solopreneurs:
- Am I creating this purely for SEO traffic?
- Am I adding anything no one else on the internet has?
If you answer yes to the first and no to the second, the piece is dead before it goes live. Don’t publish. Don’t even finish drafting it. That is the new bar for AI content that ranks.
What does AI content that ranks actually look like?
AI content that ranks today has four things working together. Uniqueness, real experience, engagement signals, and structure. None of those four are optional anymore.
Let me walk through each one in the context of an AI marketing solopreneur, not an enterprise SEO team.
Uniqueness
This is the bedrock. If a competitor with more domain authority can publish the same idea, you lose. The only way to win is to write things that depend on your own experience, your own data, or your own opinion. (Yes, opinion still counts. Probably more than ever.)
Real experience
Show the thing you actually tested. Screenshots, prompt comparisons, what broke, the embarrassing first draft. This is the part most AI-generated marketing content cannot fake, because if you did not actually run the test, the writing reads like a Wikipedia entry. Generic. Beige. Forgettable.
Engagement signals
Time on page, comments, shares, repeat visits. Click-based marketing is dying. (Harry’s words, but I agree.) What matters now is whether people actually stay, read, share, and come back. AI content that ranks needs to be content people care about.
The good news for solopreneurs is that we can move on engagement signals faster than enterprise teams can. We can answer comments. We can reply to emails. We can update a post the same day someone points out a gap. Big content teams move once a quarter. We move once an hour. That speed is a moat.
Structure
I will go deeper on this in a minute, but the short version is that LLMs are bad at reading the middle of pages. Your TL;DR, headers, and conclusion are where the citations come from. If your structure is sloppy, you do not get cited.
How can solopreneurs write non-commodity AI content without enterprise data?
Solopreneurs can write non-commodity AI content by using the data they already have. Client wins, screenshots, A/B test results, real numbers from real campaigns. You do not need Similarweb. You need to write the things only you can see.
This is something I have been applying inside my own writing for AI content marketing, and it has changed how I plan every piece. Instead of starting with a keyword and trying to fill it with information, I start with a thing I actually did and work backwards into the keyword.
A few practical translations for anyone running a one-person marketing operation:
- Write about the campaign you ran last month, not the campaign theory you read about.
- Show the prompt you used, including the bits that flopped.
- Compare two AI tools by running them on the same brief, not by reading their feature pages.
- Include the screenshot. People share screenshots.
The pattern is consistent. Personal experience scales worse than commodity content, but it ranks better. That is a fair trade in 2026.
One more reframe. Treat every AI marketing post you publish as a sample of what working with you looks like, even if you never sell anything from the page. People are reading to figure out if you actually know what you are talking about. Commodity content fails that test on contact.
What about structure and the “lost in the middle” problem?
LLMs are bad at citing content from the middle of a page, so the front and end of your post are doing most of the SEO work. This is called the “lost in the middle” effect, and it is one of the most underrated structural insights for AI content that ranks today.
The implication is uncomfortable. Most marketers were taught to bury the lede so people would keep scrolling. That was a 2019 strategy.
In 2026, AI content that ranks looks like this:
- Answer the question in the first two sentences of the post.
- Lead each H2 with the direct answer, then expand.
- Include a TL;DR or summary near the top.
- Close with a clear concluding pattern, not a soft fade.
- Use lists, tables, and short paragraphs.
If you want to go deeper on this, I wrote about how to structure pages for AI search inside AI SEO Strategy. The TL;DR is that scannable content with strong entity placement gets cited; long flowing prose gets summarized away.
There is also a tools question worth flagging. I covered the eight AI writing tools I actually use in Eight AI Writing Tools for Marketers in 2026, and the ones that pass the non-commodity test all have something in common. They get out of the way and let you write in your own voice. The ones that do not pass produce beige.
The Bigger Pattern Emerging in AI Marketing Content
The moat is no longer keywords. It is perspective.
For ten years, SEO rewarded the people who could ship the most pages on the most keywords with the most acceptable level of quality. AI broke that game. Anyone can now ship a thousand pages by Friday. The question is whether any of those pages deserve to exist.
This is honestly the most exciting moment for marketing writers I have seen in years. The bar is rising, not falling. Generic AI content will keep flooding the internet, but it is going to land in the same place every other content arbitrage opportunity has landed eventually. Worthless.
AI content that ranks in 2026 is going to come from people who use AI to write faster but refuse to let AI think for them. Solopreneurs are well positioned for this, because we already write from experience. The shift is just naming the thing we were already doing.
(And honestly, the people producing the most commodity AI marketing content right now are large agencies trying to scale. They are the ones with the most to lose.)
If your AI content is invisible, the answer is not to publish more of it. The answer is to publish less, write what only you can write, and let everyone else compete in the summary box.
– Daniel Midson-Short