The #SEOchat on Twitter was run by “Brian Dordevic.” These tweets are jam-packed with information, yet you can miss them because of how swiftly they emerge and go.
I don’t think it was as much adapting and changing in the immediate but to me this was a big step forward in google getting more specific in what it was able to serve #seochat
Mordy Oberstein
A1 Shifted to keyword planning based on what content might rank higher with AI determining rank.
More concise, less meaningless words for long tail keyword ideas
Sweepsify
A1: There isn’t really anything to optimize as it aims to emulate our natural language.
I was still “fresh” so it didn’t affect at all my performance.
For sure, it helped a lot with ranking compared to pre-BERT.
Marco Giordano
A2 Not really affected. Have seen others getting decimated, but hard to say why, since we also have HCU rolling out. Some of these sites had a lot of easily detectable AI content. #SEOChat
Testing Stuff
I’ve seen a lot of back and forth volatility but nothing that seems long term (I think it was the HCU not the LSU) #SEOchat
Mordy Oberstein
A2: Zero effect, didn’t check in detail as I have many borrowed properties.
I think it was the usual scare strategy!
Marco Giordano
A3: Give us a multiple choice option for this one ??
I think that amplification via news outlets, trusted industry sources, and other Google properties such as Scholars will play a bigger factor in determing link quality.
Sweepsify
I have always and will always not be fond of link building #SEOchat
Mordy Oberstein
A4: Semantic Clustering is one example cc
@LeeFootSEO Entity Extraction via pertained models is my most common use case though. Nowadays, I mostly use Analytics and rule-based approaches instead because they are 9/10 the best choices for many problems.
Marco Giordano
A4 Does training your content tools count?
Using machine learning to generate meta descriptions atm Working on many more use cases
Sweepsify
A4: Keyword research and XML sitemap audits In a nuthshell, I’m using it for descriptive analysis and especially pure analytics
Simone De Palma
Language structure. It should be “easy” (nothing is easy) to profile language structure. I think they already do for things like “first hand experience when reviewing a product” #SEOchat
Mordy Oberstein
A5: syntax as already said by Mordy.
Some patterns are not natural for most of us or can be too “off”.
There are other alternatives but I think this one is among the quickest.
The question is: will they really do it? We don’t know.
Marco Giordano
A5: A startup is already claiming they can do it https://siliconangle.com/2022/12/16/startup-says-can-reliably-detect-ai-generated-content/… I think Google will monitor these startups and then deploy their own internal tools
Sweepsify
That is all for the section of SEO CHAT
Jeremy’s Take: AI Search Isn’t SEO With a New Coat of Paint
The biggest mistake I see with ‘AI and SEO’ is treating ChatGPT and Claude like Google with a chatbot bolted on. They’re consensus engines — they surface what the broader ecosystem says about you, not what you stuffed into a title tag. Keep optimizing the old way and you’ll keep ranking the old way: in Google, and nowhere else.
AI search is not SEO 2.0. If you keep treating it like just keywords and rankings, you’ll get keywords and rankings — in Google. If you’re trying to show up in ChatGPT and Claude and Perplexity, you’re not going to see it.
Brittany Trafis, Unscripted SEO Podcast
Different engine, different game. Brittany Trafis explains what actually moves the needle in AI search in “AI Search Is Not SEO 2.0.”
Get cited by AI, not just crawled by Google.
Schema markup hands the models the structured facts they need to understand and recommend you. We make your entity machine-readable for the AI-search era.