GEO researchers to follow: Awesome GEO reading list authors
To start on the Awesome GEO reading list, read Pranjal Aggarwal for the GEO benchmark, Mahe Chen for what drives AI citations, and Haofei Xu for Google AI Overviews. Dmitry (Dee) Kargaev, who maintains Awesome GEO and wrote two of the listed papers, wrote this page; each person is listed for a paper they authored.
The Awesome GEO reading list holds 32 academic papers and was checked on 11 November 2026. Within each group below, entries run by year and then in the list’s own order.
Benchmarks and methods
These authors are listed on the Awesome GEO reading list for benchmarks and optimization methods for generative engines.
Pranjal Aggarwal
“GEO: Generative Engine Optimization” (2024, KDD) is the paper the reading list calls the foundational GEO paper. It introduces the GEO-BENCH benchmark and reports up to 40% visibility gains from adding statistics and citations in its benchmark setting.
Haritz Puerto
“C-SEO Bench” (2025, NeurIPS D&B) is a benchmark that tests whether conversational SEO methods work across QA and product recommendation settings.
Puneet S. Bagga
“E-GEO: GEO for E-Commerce” (2025, NeurIPS 2026 E&D) is described by the reading list as the first e-commerce GEO benchmark, with 7,000+ consumer product queries with rich intent and constraints.
Yujiang Wu
“AutoGEO” (2026, ICLR) learns generative-engine preferences and extracts content optimization rules.
Sunghwan Kim
“SAGEO Arena” (2026, KDD) is a realistic search-augmented GEO evaluation across retrieval, reranking, generation and structured web-document signals.
How engines cite
The authors here are listed for studies of what generative engines cite and what shapes their answers.
Alexander Wan
“What Evidence Do Language Models Find Convincing?” (2024, ACL) studies which evidence characteristics influence RAG-style language model answers.
Haofei Xu
“Measuring Google AI Overviews” (2026, IMC) reports that AI Overviews appeared on 13.7% of 55,393 queries, and on 64.7% of question-form queries. It also reports that about 30% of cited domains were not in first-page results and that 11.0% of claims were unsupported.
Mahe Chen and Nick Koudas
“What Drives AI Citations?” (2026) works from 158,099 answers with 468,037 citations. In its finding, the engine’s own retrieval explains far more variation than page properties do, 32% against 2.8%.
Yunxuan Jiang
“Understanding ChatGPT Citations” (2026) traces query rewriting, retrieval and citation across ChatGPT 5.2 to 5.5 settings. It finds that pages retrieved more often are cited more often.
Risks and critiques
Papers on manipulation, publisher traffic and the limits of earlier GEO results make up this group of the Awesome GEO reading list.
Samuel Pfrommer
“Ranking Manipulation for Conversational Search Engines” (2024, EMNLP) shows how conversational search ranking can be manipulated. The reading list marks it as a useful safety caveat for GEO tactics.
Stephanie T. Wang
“AI in Search Reduces Publisher Referrals” (2026) is a preregistered field experiment on Google Search with N=1,100. It found that removing AI Overviews and AI Mode raised clicks to publishers, and that an AI Mode-only experience lowered clicks and user trust.
Olivier Martinez
“Critical Survey of GEO (2023–2026)” (2026) reviews 45 studies. It reports that topical relevance and context position are the most reproducible levers, and that the foundational paper’s gains hold only when a source is already in the model’s context.
Frameworks
This group of the Awesome GEO reading list holds frameworks for generative engine optimization.
Beining Wu
“MAGEO” (2026, ACL Findings) is a multi-agent GEO framework that learns reusable, engine-specific optimization strategies with fidelity-aware evaluation.
Dmitry Kargaev
“The SEO-to-GEO Gap” (2026, SSRN working paper) compares traditional SEO and GEO ranking-factor evidence. One of its provisional findings is that visibility in generative search appears more tightly associated with entity and brand prominence than with classic link-proxy measures, and it argues for a reweighting model rather than SEO replacement. “SEEN: A Four-Layer Framework for GEO” (2026, SSRN working paper) groups GEO work into four layers: Structure, Evidence, Entity and Notability. The paper does not claim that the framework guarantees citations, recommendations, rankings or traffic.
Practitioners named by other expert lists
This list follows publications, so it misses people who do not publish. Four other expert lists name practitioners. Each description below is what that list says, and each list includes its own author.
Ahrefs’ list of 25 people, written by Ryan Law, counts original AI search studies. It says Kevin Indig publishes his studies at growth-memo.com and that Xibeijia Guan’s studies cover how much AI Overviews cut clicks. It describes Louise Linehan as a content marketer at Ahrefs who leads much of its AI search research writing.
Metehan Yesilyurt’s article names Malte Landwehr, Tomek Rudzki and David Konitzny alongside himself as people who, it says, reverse-engineer how generative engines retrieve, rank and cite sources. It says Dan Petrovic’s research covers how AI search grounding works across Google, OpenAI and Anthropic.
Josh Blyskal’s list says Tomek Rudzki published a query fan-out study covering five million hidden searches across ChatGPT, Perplexity and Grok. It also says Jan Ehrlinspiel led a study of nearly 200,000 AI responses on how rank inside frequently cited listicles affects brand visibility.
Olaf Kopp’s page describes Michael King as founder and CEO of iPullRank and calls “relevance engineering” his most important contribution to GEO. It describes Hanns Kronenberg as Head of SEO at chefkoch.de, publishing at the interface between information retrieval and new search technologies. It credits Kopp himself with a database of search engine patents and research papers.
Entries at a glance
| Entry | Year | What it shows |
|---|---|---|
| Pranjal Aggarwal, GEO: Generative Engine Optimization | 2024 | GEO-BENCH benchmark; up to 40% visibility gains from statistics and citations in its benchmark setting |
| Haritz Puerto, C-SEO Bench | 2025 | Benchmark of conversational SEO methods |
| Puneet S. Bagga, E-GEO | 2025 | E-commerce GEO benchmark with 7,000+ product queries |
| Yujiang Wu, AutoGEO | 2026 | Learns engine preferences and extracts optimization rules |
| Sunghwan Kim, SAGEO Arena | 2026 | Search-augmented GEO evaluation |
| Alexander Wan, What Evidence Do Language Models Find Convincing? | 2024 | Which evidence characteristics influence RAG-style answers |
| Haofei Xu, Measuring Google AI Overviews | 2026 | AI Overviews on 13.7% of 55,393 queries |
| Mahe Chen and Nick Koudas, What Drives AI Citations? | 2026 | Engine retrieval explains more variation than page properties |
| Yunxuan Jiang, Understanding ChatGPT Citations | 2026 | Pages retrieved more often are cited more often |
| Samuel Pfrommer, Ranking Manipulation for Conversational Search Engines | 2024 | Conversational search ranking can be manipulated |
| Stephanie T. Wang, AI in Search Reduces Publisher Referrals | 2026 | Removing AI Overviews and AI Mode raised publisher clicks |
| Olivier Martinez, Critical Survey of GEO | 2026 | Review of 45 studies on reproducible levers |
| Beining Wu, MAGEO | 2026 | Multi-agent framework with engine-specific strategies |
| Dmitry Kargaev, The SEO-to-GEO Gap and SEEN | 2026 | Provisional reweighting model; four-layer framework |