Alex Moss
Key takeaways
- 5:04Schema markup has nearly tripled from under 300 types in 2011 to 799 today, but growth has plateaued—Google removed seven types in the last six weeks, suggesting the system has reached its optimized ceiling.
- 7:07AI systems require structured content beyond JSON-LD markup: concise copy, clear hierarchy, proper heading order, and tokenization awareness—LLMs ignore marketing fluff and demand direct, honest answers.
- 34:34Product variant schema is critical for e-commerce; if you offer shoes in seven sizes, all seven must be properly marked up as distinct product variants or AI systems may not surface your page for size-specific queries.
- 17:15Measuring AI-driven results is fragmented across emerging tools; track AI Overview source citations with URL hash highlights, monitor ChatGPT/Claude interactions via server logs, and use natural language search tools to validate how AI systems interpret your content.
- 23:25Black hat SEO tactics are becoming redundant because LLMs detect manipulation patterns instantly with no algorithmic delay; instead, focus on authority, aged mentions, and transparent, concise content that serves both users and AI.
Questions Alex Moss answered
Alex fell into SEO as a teenager building fan sites and doing accidental black hat tactics, then progressed into WordPress around 2009 where he became fascinated with structured data and entity relationships, eventually joining Yoast to work on schema implementation at scale.
AI-ready schema: Feeding Google’s answer engines with structured data3:03
Schema started as simple SERP enhancements (recipes, reviews) and grew from under 300 types in 2011 to 799 today. Growth has plateaued with seven types recently removed, suggesting optimization has peaked. Schema now feeds AI systems and LLMs beyond just traditional search.
AI-ready schema: Feeding Google’s answer engines with structured data4:04
AI-ready goes beyond structured data markup to include structured content: concise, hierarchical information that LLMs can tokenize efficiently. It requires thinking about how content flows and is presented, treating LLM ingestion differently than traditional crawler interpretation.
AI-ready schema: Feeding Google’s answer engines with structured data7:07
LLMs ingest content differently than crawlers, prioritizing conciseness and clarity over marketing fluff. Content structure—information order, concept hierarchy, and relationships—matters significantly. Using tokenization tools to understand how LLMs parse content is now essential for optimization.
AI-ready schema: Feeding Google’s answer engines with structured data9:08
Google's AI Overviews extract and synthesize content, then cite it with source links that jump to contextually highlighted sections (marked with hashtags). Tracking these highlighted pages via server logs and GA4 allows measurement of which content pieces Google's AI finds most useful.
AI-ready schema: Feeding Google’s answer engines with structured data17:15
In Alex Moss's words
“I didn't realize I was doing SEO when I was a teenager. I used to build websites like I had a South Park fan website and I was doing black hat SEO without realizing that that was what I was doing.”
AI-ready schema: Feeding Google’s answer engines with structured data2:03
“LLMs don't care about any of that. They just want get to the point. Get what do you do? And be clear and concise and honest and transparent because LLMs don't care about content marketing fluff around what the actual point and the intent of it is.”
AI-ready schema: Feeding Google’s answer engines with structured data10:08
“How do you remove that zero-click issue? How do you make that click happen? I guess that's just like in a marketing way of just giving someone something to wet their appetite that might require more investigation in order to come to that your site.”
AI-ready schema: Feeding Google’s answer engines with structured data13:11
“Black hats have hated that all that time. And now I go back when from an agency point of view, I've always said unbranded mentions are good anyway. Now LLM will consider that as aged authority, which now some agencies might be going, crap, I've been doing it one way all this time to try and game an algorithm where now it's gone beyond anything I can control.”
AI-ready schema: Feeding Google’s answer engines with structured data24:26
“If ChatGPT or OpenAI released a browser tomorrow, I think a lot of people would install it for sure.”
AI-ready schema: Feeding Google’s answer engines with structured data30:29