Blog
Blog
All posts · page 8.
145 posts
Does structure alone raise AI citations? The evidence
An empirical study that raised AI search citation rates 17.3% by tuning only document structure (headings, paragraph length, emphasis), plus how engines differ.
llms.txt Setup Guide: robots.txt, Templates, and Impact
Unlike robots.txt, llms.txt acts as a content signpost. Six steps and a template, real data on AI citation impact, and when to make it a priority.
llms.txt Is a Signpost for AI Agents, Not a Ranking Button
llms.txt doesn't affect rankings. It's a signpost that helps AI agents understand and act on your site. What Lighthouse checks, and what to expect and prepare.
How Korean SEO differs from the US: Naver, Google, and AI
Korean search runs on three channels: Naver, Google, and AI search. How Naver and Google SEO differ, what to do first on a small team, and one-content tactics.
Intent-based keyword research in 6 steps
Intent-based keyword research sorts search intent into four types first. Volume vs. intent, whether to chase low-volume long-tail terms, and a 6-step process.
5 signals for keyword expansion beyond volume and difficulty
Volume and difficulty aren't enough to pick keywords. Score candidates on buyer proximity, performance gap, topic depth, AI citability, and search realism.
How to get cited by AI: 7 practical steps
Getting cited by AI takes definition sentences, structured data, inline citations, and trust signals, in order. Where to start, and what to do if it fails.
SAGEO Arena: edit only the body and you drop out of search
SAGEO Arena re-evaluates GEO across retrieval, reranking, and generation. Why title and meta fields beat body text in AI search, and how to survive reranking.
Cut LLM Misclassification with Hierarchical Classification
Let AI judge concrete categories and let code map them to abstract parent concepts, and LLM misclassification drops sharply. Why direct abstract labels fail.
Design principles for a GEO data AI agent
Design principles for a GEO data AI agent: text parsing vs. structured tool calls, blocking irreversible actions, and why skipping evals is risky.