OD
Official documentation
Product and platform docs
Useful for definitions, interfaces and first-party statements. Strong on specificity; not independent evidence.
PrimaryOpen access
Generative engine optimization · AI search evidence
The source atlas for AI search.
See which sources AI answers draw from, inspect the evidence, and track how those patterns change.
15 source profiles10 observed questions1 open benchmarkPreview scope
Start with the evidence
Trace the answer back to its sources.
SEOGEO connects generative engine optimization questions, observed AI responses, cited source URLs and the method used to interpret them.
01Source Atlas
Who shapes the answer?
Explore the institutions, publishers and primary sources behind AI search results.
Browse sources02EvidenceWhat was actually cited?
Read dated observations with the exact question, answer context and cited URLs.
Open the evidence log03ResearchWhat changes over time?
Follow repeatable studies of source mix, citation behavior and answer drift.
Read the researchLatest research · Preview
How AI answers divide their sources.
A repeatable view of source mix across questions, products and fresh conversations.
What a source mix can tell you before you count citations.
We record the full question, a dated answer snapshot and the URLs that actually appeared. The result is a source map grounded in observations, not a prediction of rank.
- Every observation carries its method version.
- Dates and limitations stay visible.
- Mentions are separated from linked citations.
Illustrative source mixPreview data
Official docsResearchEditorialCommunity
Visual placeholder for layout review. Replace with sourced results before publication.
Source Atlas
Not every source earns trust the same way.
Find a source by its role, evidence value, access model and commercial relationship.
SR
Standards and research
Papers and technical standards
Useful for tested claims, methodology and stable definitions. Always check the sample and publication date.
PrimaryMethod-led
IP
Independent publication
Editorial and industry data
Useful for interpretation and market context. Separate reported fact from editorial opinion.
SecondaryContext
PE
Practitioner evidence
Experts, communities and case work
Useful for first-hand experience and emerging practice. Treat it as a lead, not automatic proof.
ExperientialVariable
Evidence log
A citation is more than a mention.
An answer can name a brand, restate its claim, or link directly to its page. SEOGEO keeps those outcomes separate.
Cited
Question 1
4 cited sourcesHow do AI search engines select their sources?
A source URL appears in the answer.
Mentioned only
Question 2
3 named sourcesWhich search products publish retrieval guidance?
The brand appears without a linked source.
Not observed
Question 3
0 target citationsHow should teams measure AI citations?
No target source appears in this run.
The method
A conclusion is only as useful as its method.
Every source record and citation observation shows what was checked, when it was checked, and what the evidence cannot prove.
Read the method01
Show the question.
Record the prompt, language and intent without rewriting it.
02
Keep the timestamp.
AI answers move. A dated observation is more useful than a timeless claim.
03
Separate fact from inference.
A citation can support one conclusion while leaving another unresolved.
Guides
Read the evidence correctly.
Practical notes on what source records can show, what they cannot, and where interpretation begins.
GEO guide 01
What is generative engine optimization?
The meaning and practical scope of GEO, including what it can improve and what it cannot guarantee.
Read note 9 min readAI search guide 02How to get cited by AI search
What teams can control when building a page that AI systems can retrieve, understand and verify.
Read note 8 min readMeasurement guide 03How to measure AI citations
Build a repeatable observation set without confusing links, mentions, rankings or guarantees.
Read note 9 min readGEO and AI search questions
What people ask before trusting an AI answer
Short answers with links to the evidence, definitions and limits behind them.
What is generative engine optimization?
Generative engine optimization, or GEO, is the practice of making content easy for AI-powered search systems to retrieve, understand, verify and cite. It depends on useful original content, clear entities, accessible pages and evidence; it cannot guarantee that an AI product will cite a page.
Is GEO replacing SEO?
GEO extends SEO rather than replacing it. Crawlability, useful content, clear page structure and trustworthy signals still matter. GEO adds attention to the evidence an AI system can extract and the sources it can verify.
How do AI search engines choose which sources to cite?
Observed answers commonly combine retrieval, query relevance, source clarity and authority signals. The exact process differs by product, market and time, so SEOGEO records dated citations and treats one answer as one observation rather than a permanent preference.
Is SEO still relevant in 2026?
Yes. Search remains an important discovery surface and its technical and content foundations continue to support AI search systems. The difference is that teams also need to measure mentions and citations inside synthesized answers.
How can teams measure AI search visibility?
Use a stable question set, repeat observations, and record the product, locale, date, answer and every exposed citation. Separate linked citations from unlinked brand mentions and keep method versions attached to the dataset.