GEO
How ChatGPT Selects Brand Images in Visual Answers
A proposed framework for investigating image selection, metadata, and source attribution in AI answers.
2 min readQuestions / methods / evidence
AI Search, GEO & LLM Citation Research
Explore research notes and proposed methodologies for investigating how AI search platforms retrieve, display, and cite web content. Questions span ChatGPT Search, Google AI Overviews, source attribution, visual answers, and AI visibility measurement.
What the notebook documents
Research starts with a testable question and an inspectable method. A proposed methodology describes how to investigate it; an observation records what was seen in a stated context; a completed experiment needs its actual setup, results, and limitations. Commentary offers interpretation and should be read alongside its supporting sources.
The current published notes are a proposed image-selection protocol and a citation measurement framework. They are plans for collecting evidence, not completed experiments or proof of a platform's internal ranking or citation rules. No sample results are asserted.
Questions to investigate
AI citations
How can mentions, linked citations, and recommendations be recorded separately?
Research question · no result claimedVisual answers
Which images and source links are visible, and how can observations be logged consistently?
Research question · no result claimedSource attribution
What information is available on a cited page, and what remains uncertain about selection?
Research question · no result claimedCrawling and discovery
Can a published page be accessed, read, and inspected by a crawler?
Research question · no result claimedPublished research notes
Research note · By Shanmugaraj Y
Research note · By Shanmugaraj Y
AEO
From Rankings to Citations: Measuring AI Search Visibility
A practical measurement plan for mentions, citations, and brand visibility across AI platforms.
2 min read