#aio
"Our director of content and SEO, Elisa Gabbert, sees things differently. She has argued that optimizing solely for AI search is misguided—and not just because the click-through rates are so low. She also noticed that the AI Overviews she saw for similar queries cited different sites every time. If there’s no consistency in what Google chooses to pull into the AIO, how can you even figure out what they’re looking for? What assurance would you have that a citation wouldn’t disappear the next day?
(...)
One thing that stood out to me while conducting this experiment was an interesting formula Google seemed to follow for AI Overviews. Results for each search reliably included:
- A few industry-specific sites
- A community platform result
- A YouTube video (or two)
- One to two education, government, or nonprofit sites
- Often a social media post or two (typically TikTok or Facebook)
So you are seeing some variation in the types of content that Google is pulling into results, but they almost always fall into one of these categories. And if your site doesn’t fall into any of these categories, you might be out of luck, regardless of how useful or optimized your content is."
Interesting analysis!
https://www.wordstream.com/blog/ai-overviews-stability-study
Früher (vor 2022): Suchmaschinenoptimierung für Internetseiten.
Heute (seitdem KI-Agent:innen für Web-Recherchen eingesetzt werden):
"Optimierung Ihrer Inhalte, damit KI-Agents Ihre Inhalte als relevant einschätzen und sie für ihre Ausgaben in Betracht ziehen und referenzieren."
(sinngemäß)
Dafür notwendig:
AIO = AI Optimization
AXO = Agent Experience Optimization
GEO = Generative Dngine Optimization
agents.md
llms.txt
usw.
"Wir prüfen, ob die Texte auf Ihrer Internetseite für KI-Agents optimiert sind."
Das wird wohl der neue Standard.