Part one of the twelve-part series AI Content: 10 Strategies That Work. A new part every Friday.
Most guides to producing content with artificial intelligence that are still circulating today were written in 2023. Since then, essentially everything that matters has changed. The models are somewhere else, the prices are somewhere else, and above all the route a finished text takes to reach a reader is somewhere else.
Three numbers instead of an argument
Between January and April 2026, 68 % of Google searches in the US ended without a single click. Someone asked a question, got an answer and went no further. For sites that live on search traffic, that is not a footnote in a methodology, it is a hole in the business model.
The second number measures the same thing at one specific position. According to Ahrefs, the click-through rate of the first organic result for queries with an AI Overview fell by 58 % compared with December 2023. Ranking first still pays off, just far less than it used to, and the gap between first and second place loses its drama when both sit below an answer the reader has already read.
The third number comes from Pew Research Center and covers almost 69,000 real searches. When an AI summary appears above the results, people click through to a traditional link in only 8 % of cases instead of the usual 15. Roughly half of all clicks vanish simply because something is displayed above the results.
Put differently: content is cheaper and faster to produce than ever before, while the route it used to take to the reader has largely closed. „Publish a lot and publish often“ sounded reasonable in 2023. Today it is a way to burn money.
Why this changes the brief, not just the tactics
The usual reaction to these numbers is to go looking for a replacement tactic. When one thing stops working, people try another, usually whatever is being talked about at conferences that season. The problem is that the tactics did not change. The economics of the whole discipline did.
The cost of producing a text has fallen close to zero. The cost of a reader’s attention has stayed exactly where it was, because the reader is the one paying it, not you. The distance between those two curves is the whole story of the past three years, and it explains why adding more articles stopped working as a growth lever.
That leads to an uncomfortable but liberating conclusion. Being able to produce an article is no longer a competitive advantage, because everyone can. The advantage is only in what you put into the text that is not available anywhere else: your own numbers, your own experience, a specific name, a specific date and an opinion someone is willing to stand behind.
What this series is about
It is not about how to squeeze an article out of a language model. Social media is full of „write me a blog post“ prompt guides and their value matches how easily they can be copied.
It is about what to do with the fact that an article has no value on its own any more, because an unlimited number of them exists. We will go through ten strategies, from the publishing plan through briefing the model and fact-checking to measurement, distribution and the legal obligations that have applied across the EU since this August.
Every strategy rests on documented measurements and sources you can cite, not on impressions. Where a misconception has taken root in the industry, we name it along with where it came from and who wrote it down first. With two of those myths it turns out nobody ever documented them, they simply get copied onward.
What is coming in the next parts
Next Friday we start with the publishing plan and with a piece of advice you will find in nine out of ten SEO handbooks and which was never true. Then come prompts and fact-checking, the part without which everything else falls apart, and personalisation, which does not mean dropping a name and an industry into the text.
The middle of the series belongs to visibility in a year when people stopped clicking, to the division of labour between a human and a model (including why AI text detectors do not work), and to images. Images are where we take apart one of the most durable marketing myths there is.
The closing parts are practical and boring in the best sense of the word: measurement that actually tells you something, breaking a single text down into several channels, law and ethics as they stand in August 2026, and finally the system that holds all of it together without swallowing your entire week.
A note on shelf life
The series is written for 2026 and it will age. Tool prices change from month to month, legislation from year to year, and Google’s recommendations according to whatever it happens to be testing.
For every claim that rests on a measurement or a legal provision you will find a source in the text itself, and all of them together in the final part. If a number surprises you, go to the primary source. That, incidentally, is exactly the procedure we recommend for working with AI output.
Check prices and the current wording of legislation before any decision with a financial impact. Not because they are wrong here, but because they change faster than a new version of the text can come out. The series is based on an ebook we prepared for Mediatoring.com clients.
Next Friday: A publishing plan that assumes volume does not work.
The whole series: AI Content
1. Why most advice about AI content no longer holds · 2. A publishing plan that assumes volume does not work · 3. The prompts that make a difference, and the fact-checking · 4. Personalisation that is more than a different salutation · 5. Visibility in the year people stopped clicking · 6. Where AI ends and a human begins · 7. Images, and how to spot a marketing myth · 8. Measurement that tells you something · 9. One piece of content, several channels, without copy-paste · 10. Law and ethics, specifically, as they stand in August 2026 · 11. Build a system, not a pile of articles · 12. What to take from the whole series