Where the Term Comes From
GEO was formalized in November 2023 by researchers at Princeton, the Allen Institute for AI, and IIT in the paper 'GEO: Generative Engine Optimization' (Aggarwal et al., arXiv:2311.09735). They introduced GEO-bench — a benchmark of 10,000 queries — and tested which content tactics increased source visibility inside LLM-generated answers.
Headline finding: targeted content optimizations boosted source visibility in generative engine responses by up to about 40%.
Source: arxiv.org/abs/2311.09735.
SEO vs GEO — Same Goal, Different Game
SEO is about ranking on a list of links. GEO is about being inside the answer. The signals overlap (crawlability, E-E-A-T, topical authority, internal linking, fresh content), but the optimizations diverge.
GEO rewards passage-level extractability. The LLM grabs a sentence or paragraph and quotes it — so the content needs to be quotable on its own, with concrete claims and clear attribution. Backlinks matter less. Being directly quoted matters more.
Source: schema.org documentation, Google Search Central, arXiv:2311.09735.
Why It Matters Now
Google AI Overviews launched in May 2024 and reached an estimated 1.5 billion monthly users by Q1 2025 (Alphabet earnings). ChatGPT Search rolled out in October 2024 and reached free users in February 2025. ChatGPT itself reports 300M+ weekly active users (OpenAI, late 2024).
Studies from BrightEdge and seoClarity put AI Overviews on roughly 13–20% of US search queries in 2024, varying by vertical. That share is climbing, not flat.
Source: Alphabet Q1 2025 earnings, OpenAI press releases.
Who Gets Cited and Why
Wikipedia, Reddit, YouTube transcripts, Stack Overflow, GitHub README files, and long-form expert blogs (Stratechery, Ahrefs, NerdWallet) get cited disproportionately. The pattern: structured Q&A or canonical-answer format, neutral tone, heavy use of stats, quotes, and clear headings.
Reddit specifically benefits from a Google licensing deal (Reuters, February 22, 2024) that pipes its content into AI Overviews. Off-site citations from Reddit threads now influence what AI engines say about your business.
The llms.txt Proposal
In September 2024, Jeremy Howard (Answer.AI) proposed `/llms.txt` — a markdown file at the root of a site giving LLMs a curated map of key pages. It is voluntary and not yet honored by every engine, but adopting it costs you nothing and signals intent.
Source: llmstxt.org.