GEO.FanResources
A curated index of GEO learning resources — academic papers, industry reports, the Chinese AI ecosystem map, tools, and courses.
Machine translation
TL;DR: This page is a curated resource index for GEO practitioners — grouped by academic papers, industry reports, the Chinese AI ecosystem, tools, and courses. Every entry links to its original source for deeper reading.
This page draws heavily on the community-maintained LLM-X-Factorer/awesome-geo-cn (Chinese GEO resource list) as well as multiple 2025–2026 industry reports. Verify external links yourself — link rot in the AI / search space is fast.
Reading order (this site)
If you're new to GEO, we suggest reading in this order:
- What is GEO — concept introduction
- GEO vs SEO — key differences from traditional SEO
- Core Concepts — scoring dimensions + Chinese AI ecosystem
- Quick Start — your first GEO check
- How Generative Engines Work — the technical principles
- llms.txt and AI Crawlers — technical foundation
- Content Strategy Best Practices — long-term content operations
Academic papers
GEO isn't marketing jargon — there's checkable academic research behind it. These are the foundational and representative papers in the field.
Foundational papers
| Paper | Source | Key findings |
|---|---|---|
| GEO: Generative Engine Optimization | KDD 2024 · IIT Delhi + Princeton | Quantified AI visibility lift: Cite Credible Sources +43%, Add Statistics +33%, Answer-first structure +18%, technical terminology +11%, Keyword Stuffing is a negative effect |
| GEO and the Future of Search | KDD 2024 Workshop | GEO conceptual framework and evaluation methodology |
2025+ follow-up research
| Paper | Source | Highlight |
|---|---|---|
| Generative Engine Optimization: How to Dominate AI Search | 2025 | AI search systematically prefers earned media (third-party coverage, industry media) over brand-owned content |
| Self-Promotion in LLM Recommendations | Friedler et al. 2025 | LLMs exhibit self-promotion bias when recommending AI products: vendor model rankings average +0.2 |
| LLMs are Biased Evaluators But Not Biased for Fact-Centric Contexts | ACL 2025 Findings | Bias hierarchy in RAG scenarios: factual > order > self-preference |
| C-SEO Bench: Does Conversational SEO Work? | NeurIPS 2025 | Under strict controls, most GEO methods are basically ineffective; effects cancel out when multiple parties use GEO simultaneously |
| AutoGEO: What Generative Search Engines Like (code) | ICLR 2026 | Uses GRPO to automatically learn generative engines' content preferences, proposes a "collaborative" optimization framework |
| MAGEO: Multi-Agent GEO via Reusable Strategy Learning | ACL 2026 | Multi-agent collaborative learning of GEO strategies, codifying experience into reusable skills |
| E-GEO: A Testbed for GEO in E-Commerce | 2025 | 7,000+ e-commerce query benchmark evaluating 15 rewriting heuristics |
| A Survey on Retrieval-Augmented Generation Optimization for LLMs | Institute of Computing Technology, Chinese Academy of Sciences, Chinese Journal of Computers 2026 | First-hand Chinese RAG survey: query rewriting, retrieval augmentation, knowledge injection, citation generation |
Adversarial and citation-mechanism research
| Paper | Source | Highlight |
|---|---|---|
| ConflictingQA | 2024 | LLMs prefer "relevance" over "academic tone" |
| ConflictBank | 2024 | 7.4M claim-evidence pairs, external content vs. training data conflict |
| GASLITE: SEO Attacks on Dense Retrieval | 2024 | 0.0001% corpus pollution can hijack top-10 retrieval |
| Adversarial SEO for LLMs | 2024 | Hidden text can boost brand mentions in AI answers by 2.5× |
| Ranking Manipulation for Conversational Search | 2024 | Prompt injection manipulates conversational search rankings |
| Dynamics of Adversarial Attacks on LLM-Based Search | 2025 | Game-theoretic modeling of black-hat vs. white-hat GEO |
Industry reports
Chinese reports
| Resource | Publisher | Notes |
|---|---|---|
| SuperCLUE-AISearch | SuperCLUE | Chinese AI search benchmark, updated monthly |
| 2026 GEO Industry Research Report | iResearch 2026 | Chinese GEO industry report: definitions, common misconceptions, cases, market sizing |
| GEO White Paper 2026 | CEIBS 2026 | Chinese GEO white paper from an academic institution |
| AI Search Product Evaluation 2025 | IDC China 2025/07 | Scenario-based comparison of Baidu / Quark / Doubao / DeepSeek |
| Five Trends of China's Generative AI Market in 2025 | Roland Berger 2025 | AI agents, multimodal, hardware integration trends |
English reports
| Resource | Publisher | Key findings |
|---|---|---|
| AI Search Visits Surging in 2025 | BrightEdge 2025/09 | Fortune 100 measurements: AI search has double-digit monthly growth but still <1% of total traffic |
| AI Overviews One Year Review | BrightEdge 2025/05 | One-year review of Google AI Overviews: search volume +49%, CTR -30% |
| Platform Citation Preferences | Hashmeta 2025/01 | Cross-platform preferences across 6 major AI platforms / 15,000+ citations / 3,400+ queries |
| LLM Citation Study by Industry | Writesonic 2025/11 | LLM citation patterns across industries; citation overlap between GPT versions is only 7% |
Key industry data (with sources)
| Data point | Source |
|---|---|
| ChatGPT 891M MAU, 17.6% of search | SparkToro 2025 |
| Google holds 77.9% of search | SparkToro 2025 |
| 60% of Google searches don't produce a click | SparkToro 2025 |
| AI search traffic YoY growth 527% | BrightEdge 2025 |
| SEO and GEO ranking overlap is only 12% | Ahrefs 2025 |
| Citation probability when content appears on 4+ platforms ×2.8 | KDD 2024 |
| AI-cited visitor conversion is 4.4–23× that of ordinary search | BrightEdge 2025 |
| Zhihu AI citation rate 29.9% | IT Home 2026 |
| Reddit AI citation rate 40.1% | SparkToro 2025 |
Chinese AI ecosystem
For the detailed Chinese AI ecosystem map (ByteDance / Tencent / Alibaba / Baidu factions + independent products + content platforms), see Core Concepts → Generative engine landscape.
Full external index: LLM-X-Factorer/awesome-geo-cn — continuously maintained Chinese GEO resource list.
Tools
GEO monitoring (commercial)
| Tool | Description | Price |
|---|---|---|
| Ahrefs AI Overviews Tracker | Track brand citations in AI Overviews | Paid |
| Semrush AI SEO Toolkit | AI search visibility analysis | Paid |
| Profound | AI search citation monitoring | Paid |
Open-source tools
| Tool | Description |
|---|---|
| AI2HU/gego | Go-implemented GEO tool, cross-LLM brand exposure tracking, REST API |
| aircodelabs/llms-txt-generator | AI-driven llms.txt / llms-full.txt generator with MCP integration |
| apify/actor-llmstxt-generator | Apify Actor form of llms.txt generator |
| infiniflow/ragflow | Open-source RAG engine with deep document understanding (79k+ stars) |
| danishashko/geo-aeo-tracker | Local-first AI visibility dashboard covering 6 major AI engines |
| Auriti-Labs/geo-optimizer-skill | GEO audit / optimization / testing tool based on KDD 2024 research |
| LLM-X-Factorer/md2red | Markdown to Xiaohongshu (Little Red Book) image-card converter |
On-site tools
| Tool | Use |
|---|---|
| GEO Checker | Content quality scoring |
| Performance Analysis | Long-term performance monitoring |
| Competitor Analysis | Horizontal benchmarking |
Courses
| Course | Language | Description |
|---|---|---|
| SEO + GEO introductory course | Chinese | 10 weeks, 29 lessons, based on the KDD 2024 paper, covers SEO + GEO + Chinese AI platforms |
| AI SEO: Mastering GEO (Coursera) | English | Coursera GEO course |
| Wanzhihui SEOxGEO Introduction | Traditional Chinese | 61-unit video course |
Curated English resources
| Resource | Description |
|---|---|
| Search Engine Land — What is GEO | Definition and overview |
| Search Engine Land — Mastering GEO in 2026 | Complete GEO operations guide (2026) |
| Backlinko GEO Guide | Brian Dean's GEO guide |
| Kevin Indig — Growth Memo | Data-driven GEO / AI search blog (once analyzed 1.2M ChatGPT responses) |
| Can You Fake Expertise in AI Search? | Tests of expert citation preferences across 9 AI models |
External lists: luka2chat/awesome-geo, amplifying-ai/awesome-generative-engine-optimization, DavidHuji/Awesome-GEO (academic index).
Templates
Content evaluation checklist
Before publishing, check that:
- The article opens with a standalone TL;DR / quick-answer block (extractable as one snippet)
- Heading hierarchy is unique and consistent (single H1, sensible H2/H3 nesting)
- 3–5 named citations per 1,000 words, linking to specific article URLs (per KDD 2024, authoritative citations lift AI visibility by +43%)
- At least one original data point / case / diagram
- Author byline + brief credentials
- FAQ section (at least 3 Q&As)
- At least 2 JSON-LD layers (Article + FAQPage or Article + HowTo)
- SSR verification:
curl -A "GPTBot" <URL>shows the body
Prompt test set template
To verify whether your content can be found and cited by AI engines:
1. I'd like to learn about <topic>. Any resources to recommend?
2. How do I do <your core problem>?
3. What does <related sub-question> mean?
4. How can I avoid <the pain point you solve>?
5. Compare <your approach> vs. <competitor approach>.Recommend running against 5 major engines: ChatGPT, Claude, Perplexity, Google AI Overviews, Gemini. For Chinese sites also add Doubao / Kimi / ERNIE Bot / Qwen. Monthly.
Feedback and contributions
Found a broken resource link, have a better tool recommendation, or want to contribute a template? Email mail@geo.fan.
Competitor Analysis
Compare your performance against competitors in generative AI engines and find optimization opportunities specific to your position.
llms.txt and AI Crawler Configuration
Make your content discoverable, crawlable, and citable by AI engines — guide to the llms.txt standard and robots.txt configuration.