Stuzhuk Lab — Chemistry of Code
Chemistry of Code
SEO and AI search visibility
Technical SEO as the foundation, AEO for extractable answers, and GEO for measurable brand visibility in generative systems.
Tell me the goal, stack constraints, and timeline — I reply on Telegram.
Conventional SEO remains the foundation: crawlers need complete HTML, the correct canonical URL, useful internal links, and a comprehensible page structure. AEO helps search and answer systems extract a short, self-contained response. GEO asks a different question: whether a brand appears in generative answers, is described accurately, and which sources are cited.
I start with technical accessibility and an intent map, reviewing routes, SSR or prerendering, sitemap, robots, hreflang, structured data, and duplicates. Then I assess whether content provides direct answers, primary facts, authorship, dates, limitations, and meaningful links. Generic advice gains no information value from FAQPage or llms.txt.
GEO needs a reproducible query set, not three favourable screenshots. The system, region, and wording are recorded; brand mentions, citations, factual accuracy, and competitor answers are measured separately. A guaranteed position is not a valid promise because models, personalisation, and undisclosed ranking methods change.
Scope follows the problem. A technical audit is not the editing of hundreds of articles, and schema implementation is not continuous monitoring. Engineering acceptance means indexable URLs, valid markup, and accessible text; content acceptance means agreed intents without cannibalisation; GEO acceptance means a repeatable report on one query set.
FAQ
Tap a question to expand the answer.
How do SEO, AEO, and GEO differ?
SEO keeps pages crawlable and trustworthy. AEO formats content for in-search AI answers. GEO tracks brand presence in standalone chat assistants via Share of Voice.
What is AISO?
AISO (AI Search Optimization) is the technical layer: crawler access, extractable HTML, structured data, and site architecture for retrieval — the foundation under AEO and GEO.
What is Information Gain?
New semantic value versus what search and LLMs already know. Rewrites without primary data (cases, metrics) lose to sources with first-party facts.
How do you measure Share of Voice?
A pool of situational, task, and comparison prompts; 60–100 runs per prompt in target LLMs; percentage of answers mentioning the brand, plus Sentiment and factual Accuracy.
What are ghost citations?
The site appears in the source list but the brand name is absent from the generated text. Fix: embed the brand in body copy, tables, and code examples — not only meta tags.
Do we need llms.txt?
Useful as an agent-oriented site map and documentation habit. Public research and Google’s position treat it as not a proven citation booster — we implement it as part of technical hygiene, not as a SoV guarantee.
Does this work for React SPAs without SSR?
Poorly. Empty or delayed HTML hurts extraction. Prefer SSR, prerender, or hybrid rendering for pages that must appear in AI answers.
Tell me the goal, stack constraints, and timeline — I reply on Telegram.
