A few years ago, a clear 800-word article on your website was a small advantage. Most businesses in your area weren’t publishing anything. Now anyone can generate that article in seconds, and plenty do. The result is a web full of pages that say the same things in the same order.
That’s the real problem with AI-only content. It isn’t that AI wrote it. It’s that it adds nothing that wasn’t already on page one.
What Google says it rewards
Google’s guidance has been consistent: it wants content made to help people, not content made mainly to rank. Its quality guidelines describe E-E-A-T: experience, expertise, authoritativeness and trust. The first E, experience, asks whether the writer has actually done the thing they’re writing about.
Google doesn’t ban AI-written pages; it looks at whether a page is useful. But a page assembled from what’s already online struggles to show experience, because your experience is exactly what the rest of the web doesn’t have.
Why AI-only articles fall flat
- They average the web. A language model writes the most likely version of an article, which looks a lot like the pages already ranking.
- They add no new information. No job you did last week, no number from your own work, no local detail.
- They hedge. “It depends on many factors” is safe, and useless to someone deciding whether to call you.
- They can be confidently wrong. Without someone who knows the work checking them, mistakes slip through.
What first-hand experience looks like on the page
You don’t need to be a writer to add it. You need the things you’d say to a customer:
- A clear opinion: “Don’t let anyone sell you a new panel before they’ve done a load calculation.”
- A rule of thumb from your work: “Nine times out of ten, the panel you have can handle a charger.”
- A common mistake, and what it costs people.
- A short story about a real job, with the private details left out.
- Local specifics: the age of the housing, the weather, the permits.
These are the details a reader remembers and an AI answer is most likely to quote, because nobody else has them.
What we learned testing this ourselves
While building Say It SEO, we ran blind side-by-side comparisons: an article written by AI alone against one written from a phone interview with the business owner, judged by a separate reviewer that didn’t know which was which.
The first result surprised us. An interview on its own wasn’t enough: our early interview-based articles scored about the same as the AI-only ones. They had the experience, but they missed some of what the pages on page one covered.
What won was combining the two. We research the topic first, so the article covers what searchers expect. Then we interview the owner for what only they know. Then we review the draft against page one and the questions people ask, and rewrite where it falls short. Those articles came out ahead in our tests. It’s a small test, not a study, but the pattern was clear: research gives an article coverage, and experience gives it something worth reading.
How to add experience without writing for hours
- Talk instead of typing. Most people say more in five minutes of talking than in an hour of writing.
- Record your answers to customer questions. Voice notes on your phone are fine.
- Use AI for structure, not substance. Let it organise and tidy; keep your words, numbers and opinions.
- Check every fact. If it isn’t something you’d say to a customer, take it out.
That’s the whole idea behind Say It SEO. We research the topics, interview you by phone for about 15 minutes, and write the article from what you said. Your own words are highlighted, so you can see exactly what came from you. See a full example.