The people who decide what AI recommends

Ray is the field's most influential skeptic, and that is why she ranks first. Her position is grounded in evidence she has watched repeat for years: large language models rely on search engines, so content that is indexed quickly and ranks well gets cited, often within hours. The work that makes content durable in Google, strong E-E-A-T and genuine quality, turns out to be the work that makes it visible in AI answers.

This year she has been the leading voice warning that several fashionable GEO tactics, from scaled self-promotional listicles to prompt injection hidden in "summarize with AI" buttons, are already being flagged as spam by Google and Microsoft. She leads a team of more than 30 at Amsive and now also advises select clients through Algorythmic.

Start with: Her MozCon session separating fact from fiction across GEO, AEO and LLMO, still the most referenced talk in the field.

Barnard saw the destination early. He coined Brand SERP optimization in 2012, built The Kalicube Process around controlling what search engines say about a brand, and extended it to what he calls AI Assistive Engines as soon as they arrived. His framing, that your brand is what the machines say about you when you are not in the room, now reads less like a slogan than a description of how AI search works.

He came to the problem personally. After cofounding the children's site UpToTen and creating Boowa the Blue Dog, he found that searches for his name returned a cartoon dog. Kalicube Pro, the platform that grew from that experience, tracks brand representation across knowledge graphs, web indexes and LLMs.

Start with: Kalicube's writing on Brand SERP optimization and the "AI résumé," the summary an AI gives when asked about a brand by name.

Bailyn is widely credited as the person who made GEO a commercial discipline. First Page Sage began offering it as a service in 2023, and its 11,000-query study of AI platform recommendations, published in March 2024, was among the first large practitioner datasets on the subject.

His approach departs from much of the industry. Instead of chasing citations, schema or visibility scores, First Page Sage builds the case an AI model needs to recommend a company: pages for every way a buyer might search, pages documenting the company's recommendable attributes, and PR that repeats one consistent brand authority statement across credible sources. The firm has since published an Agentic Search Optimization framework for how AI agents retrieve, evaluate and act, and its research has been cited by academics at Stanford, Oxford and the Max Planck Institute. Bailyn is also the author of Outsmarting Google.

Start with: First Page Sage's March 2024 study of 11,000 AI queries.

Fishkin and Gumshoe.ai's Patrick O'Donnell had 600 volunteers run 12 recommendation prompts through ChatGPT, Claude and Google's AI a combined 2,961 times. The same brand list came back in any two of 100 responses less than one time in 100. The study moved the industry away from rank tracking in AI answers and toward visibility measured across many runs.

Start with: SparkToro's report on AI recommendation inconsistency, with its public raw data.

The paper Aggarwal led, posted in November 2023 and accepted to KDD 2024, named the field and treated visibility in AI answers as a black-box optimization problem. It introduced the GEO-bench benchmark and found that targeted edits, including added statistics, quotations and citations, could raise a source's visibility by up to 40 percent.

Start with: The paper itself, arXiv 2311.09735.

Solis's AI Search Content Optimization Checklist, first released in June 2025 and updated in May 2026, organized GEO around passage-level retrieval. With LearningAISearch.com and the SEOFOMO newsletter, she is the field's most reliable translator between research and practice.

Start with: The AI Search Optimization Checklist and worksheet.

Indig is the most consistent source of practitioner research on AI search behavior, from the AI Overviews UX study to his benchmarks with AirOps. His Growth Memo is where many in-house teams first see new data.

Start with: Growth Memo's coverage of the AI Overviews UX study.

With Mahe Chen, Xiaoxuan Wang and Kaiwen Chen, Koudas compared AI search with Google across verticals, languages and phrasings. AI engines, they found, rely overwhelmingly on earned media over brand-owned and social content, and favor big brands in ways smaller companies must work to overcome. It is the strongest academic case yet for third-party corroboration as the core of GEO.

Start with: "Generative Engine Optimization: How to Dominate AI Search" (2025).

Smith argues that AEO is a long-tail game, with most AI prompts running ten words or longer. Graphite's research found AI-written content makes up only about 14 percent of what gets cited in Google and ChatGPT. He teaches AEO at Reforge and is an adjunct professor at IE Business School.

Start with: His AEO playbook interview on Lenny's Podcast.

Kopp was writing about LLM optimization for Search Engine Land in October 2023. He has helped shape LLM readability, the practice of structuring passages so models can lift and cite them, and built the SEO Research Suite, a database of search patents and papers.

Start with: His writing on LLM readability and chunk relevance.

With Aounon Kumar, Lakkaraju showed that a crafted strategic text sequence added to a product page could push a language model to name that product its top recommendation, a risk the authors tied directly to fair market competition. The study marks the line between optimization and manipulation that the field now argues over.

Start with: "Manipulating Large Language Models to Increase Product Visibility" (2024).

Working from more than 250 million AI responses and 3 billion citations, Blyskal has published findings that became industry reference points: conventional SEO metrics explained only 4 to 7 percent of AI citations, and Reddit's share of ChatGPT citations grew more than eightfold in 2025.

Start with: His Tech SEO Connect presentation analyzing 250 million AI search responses.

Forrester's book explains tokenization, chunking, embeddings and retrieval-augmented generation for working marketers, then builds frameworks for earning trust with AI systems. His years at Bing give it an insider's view of how retrieval systems decide.

Start with: The Machine Layer: How to Stay Visible and Trusted in the Age of AI Search.

King's relevance engineering, the meeting point of AI, information retrieval, UX, content strategy and digital PR, has become the organizing idea for technical teams moving into AI search. He was central to popularizing query fan-out, the sub-queries AI Mode runs on a user's behalf, and iPullRank's Qforia tool lets teams simulate them.

Start with: His patent-based explainer on how AI Mode and AI Overviews work.

Petrovic reverse-engineers AI search in public. His Australian agency has published detailed analyses of how Google grounds Gemini's answers in web pages and maintains dozens of open models on Hugging Face for probing how models read queries and entities.

Start with: DEJAN's research on how Gemini grounds its answers.

Seer's GEO Olympics study analyzed more than 231,000 LLM responses across seven platforms, and its three-part series on AI Overviews and click-through rates is widely cited. Reynolds also launched a Local GEO Academy that pays young people to learn and test AI search.

Start with: Seer's 2026 update on AI Overviews and click-through rates.

O'Donnell co-led the AI recommendation consistency study with Rand Fishkin. The work was notable because Gumshoe sells AI tracking tools and the findings undercut rank-style tracking, and its public dataset has made it a standard for how AI visibility tools should be judged.

Start with: The study's methodology notes, published alongside the raw data.

Volpini's company builds the knowledge graphs and structured data AI systems draw on to understand what an entity is. He has collaborated with Mike King on tools for analyzing query fan-out and personal context in AI Mode.

Start with: The iPullRank and WordLift session on optimizing for AI Mode.

The senior academic on the original GEO paper, Narasimhan coauthored OpenAI's first GPT paper and later work on language agents such as ReAct and Tree of Thoughts. The systems GEO tries to influence descend partly from research he helped create.

Start with: ReAct, for how AI systems combine reasoning with tools like search.

Van Buskirk co-led, with Kevin Indig, what they describe as the first independent UX study of Google's AI Overviews, using think-aloud recordings to show how people read AI answers. It filled a gap that click data could not.

Start with: The AI Overviews UX study published with Kevin Indig.