Orada ne olduğunun farkındayım – bu yüzden neden çoğu (tümü?) Şema eklentilerinin/programlarının/uygulamalarının schema.org’da 900’ün üzerinde şema varken yalnızca 15’in altında şema kullandığını merak ediyorum. Bunu korkunç derecede yetersiz bulan başka biri var mı? Özellikle yapay zekanın şemaya dayanması onu her zamankinden daha önemli hale getiriyor.
Kasım 2, 2025
Otomatik Şema
3 Yorum
Yorum Yaz
Yorum yapabilmek için oturum açmalısınız.
“Especially with AI relying on schema, making it mire important than ever.”
I think you’re in error, or trying to sell something.
> Especially with AI relying on schema, making it more important than ever.
AI does not rely on Schame – LLMs are not search engines and do not rank content – so they cannot “prefer” or rely on it.
We encourage you to test/apply critical thinking to what you read- there is a campaign of intentiaionl disinformation about LLMs and AI SEO.
[https://www.linkedin.com/posts/markseo_seo-activity-7363511170965630984-OZtu/](https://www.linkedin.com/posts/markseo_seo-activity-7363511170965630984-OZtu/)
[https://www.reddit.com/r/SEO/comments/1m1g8tp/community_llm_seo_discussion_the_query_fan_out/](https://www.reddit.com/r/SEO/comments/1m1g8tp/community_llm_seo_discussion_the_query_fan_out/)
It’s shit, but to be fair, it’s also an absolute pain in the ass to do programmatically, tear and rage inducing from a zero insight point of view without human intervention.
I’ll give you some numbers, theoretical combinations, all entities, types, fields, nesting or not, you’ve got about 10^50 … then out of that on the low side practical application is 10^8 … and my average client has about 10^6 combinations (per page) at a start from zero…
Huge fucken numbers and why simple tools limit it to the basics. It’s also why AI have a melt down if you try to go beyond their first basic template out put… or just simply update details… loop, death…llm meltdown tantrum… can’t tell you how many time AI asked me to “Please fucken stop”.
Turns out it’s a bit of a complex problem, to do programmatically… (i have a solution), but think it through, how without intervention does a program(llm or not) select max correct entities, max types, max fields, assemble, optimise, deploy… you need a way to detect services, entities, jurisdiction legals and validation, NAP data and validation, optimal order, to nest or not, which descriptions and language, how to dectect and refer to onpage information with id or page referrals.. and more… then layered into this when it comes to products you want things to be dynamic and ideally contextually built into the page. Then you’ve got onpage citations and stuff, which is a over the top but I’ve done that to.
Anyway an out of the box tool is not going to do that. Also, why I don’t use them.
I started with an infinite drop down schema builder, but holy shit is that the most tedious job in history and requires a lot of manual research and insight. Consequently, I built the most ridiculously complex system for building optimal schema in mass… which seems so simple from the outside.