Google in 2022 to replace E-A-T with an E, Experience (first hand experience) at the top. This is not a word game. When ChatGPT, Perplexity, or Google's AI Overviews have to pick out a bunch of content to be the answer, they prefer to write the source of the physical detail that actually stepped on the pit, rather than replacing the encyclopedia with a narrative. Do you have anything in your article that says, "Only one person who has done it" and decides whether it will be taken out and quoted?
Why, AI engine prefers first-hand experience. Yon
The AI engine produces answers by lowering the risk of error. The general theory is that there is a high repetition of each other, and the same is true of any article cited, so the model tends to pick out a unique, verifiable paragraph: specific numbers, time, failure, operational steps. These details, which are more intuitive and difficult to fabricate, are tantamount to attaching a layer of credibility to the content. In turn, experience signals help the model solve the problem of belonging. If it says, "When we're auditing a B2B SaaS client, 80% of the product pages have never been quoted in AI's answer," the model knows that it has a clear origin and can be quoted; it only says, "AI visibility on the product page is important," and anyone can say it without reference.
Six identifiable signals with first-hand experience.
- More than 30 websites we've examined in the past year.
- Time and course: tell us when you did it, how long it took, how many editions you changed.
- Failures and accidents: Write where you're not going, how you found it, how you'll fix it.
- The first person called the operational visual angle: "We actually get in there" is more viable than "according to observation."
- Raw material: backstage screenshots taken by yourself, actual contrasts, not graphs.
- Refutable judgment: "This will not work in certain circumstances" means you really used it.
The common denominator of these six signals is that it is difficult to outsource it to someone who has not done it. A writer who has never carried out a GEO review has written a general comment, but has not written the details of "sitemap was found missing an entire catalogue" until the third edition. And when you read it, you're going to find out whether it's an inside or out, and the model's going to be -- that's the density of the details.
Example: rewriting an empty phrase as experienced Yon
Before rewriting
"Constructive data can help the search engine understand your content, advance the opportunity to be quoted in AI's answer. It is recommended that all pages be marked with proper Schema. It's not wrong, but it can appear in any of the same articles, and the model has no reason to pick it.
After rewriting (with first hand experience)
"We added to the price page of a SaaS client's FAQPage and Product, about the sixth week, Perplexity started directly citing their price bands in response to "the estimated amount of this tool." The real point is not to mark itself: first, let's change the price of the image from the one on the page to the plain text, because the description is empty if the model is not read. The same theme, the latter gave time, the platform, the order, and the easily ignored cause and effect — that's what can be extracted from it.
Experienced signals are not adjectives, but names and numbers. "We've got a lot of experience" and we'll replace it with what you did, what happened in the middle, what you got.
Let the experience come to light: tie the content to the author.
The experience in the article needs to be captured, and both readers and models need to know who it is. This means that the content should have a well-documented background, and that the identity should be consistent on the website: the author's page, the page, the external platform's profile, and the AI engine cross-references these signals when it judges their credibility. In practice, the author's specific project has been written in a brief that is far more effective than a string of headlines — “GEO audits that have been conducted on over 30 B2B sites”, which is an experience, and “senior digital marketing experts” are self-proclaimed.
Common false experience, readers and models can tell.
- Starting with "according to our experience," there is no concrete truth behind it.
- Compiles precise numbers, but does not give context: which client, when, how.
- Fake the image of the library as a hands-on cut-off.
- Throughout the page is a success story, none of which was unexpected, compromised or failed.
Three questions.
- Is this a story that a person who hasn't done it can write by checking the data? If you can, it's not an experience.
- After all the adjectives are removed, are there any facts, numbers or steps left?
- Reader, "Who is this, in what circumstances, what results?"
Most of the content is beyond the first question. There is also a bottom line: don't create numbers for an experienced signal, the fictional "triple-growth" goes down as soon as it is asked, and the real little numbers outnumber the pretty ones. If you want to know what empirical signals your page is missing in the eyes of the AI engine and what is not quoted at all, you can schedule a GEO consultation for 30 minutes, and we'll take your page as an example.
