An alternative to the usual GEO expert listicles: 28 digital marketing practitioners, four forms of expertise, and some observations about where Search is heading.
The problem with lists of AI Search experts
Search for “top AI Search experts” or “leading GEO specialists,” and you’ll discover an astonishing number of experts, many of whom have helpfully published lists explaining why they deserve the title. Sometimes they appear alongside colleagues. Sometimes alongside competitors. Sometimes, rather conveniently, at number one. What a coincidence.
I am exaggerating, of course. Not every expert list is self-serving, and publishing one that includes you doesn’t automatically invalidate your expertise. However, the phenomenon illustrates a problem that goes deeper than personal branding: we are becoming very good at confusing visibility with expertise (and I purposely do not enter into the topic of perverting listicles to influence AI Models).
The distinction matters particularly in AI Search, where much of what marketers believe they know is based on incomplete observations, experiments under specific conditions, proprietary datasets, and interpretations of systems whose internal workings remain undisclosed.
A practitioner publishes an investigation. Someone else explains it in a newsletter. A third person turns that explanation into a framework. Before long, that framework appears in conference presentations, agency sales materials, LinkedIn carousels, and articles naming the people supposedly pioneering the discipline. Sometimes the original researcher barely gets mentioned.
SEO has often transformed hypotheses into truths through repetition. Today, a technical observation can become an established “GEO ranking factor” before anyone tries to replicate it.
That is why I wanted a different kind of list: not the most popular personalities, not the people with the most LinkedIn followers, and certainly not consultants who have convinced ChatGPT to call them experts. Instead, I wanted to recognize people from SEO and digital marketing backgrounds who have contributed something substantive to our understanding or practice of AI Search, and allow you to examine that work yourself.
What do I mean by AI Search expertise?
AI Search is an unfortunate expression in one respect: it makes a complicated ecosystem sound like a single discipline. Google AI Overviews, Google AI Mode, ChatGPT Search, Perplexity, Gemini, and other conversational systems do not necessarily retrieve, evaluate, synthesize, or present information in the same way.
Some answers rely on live retrieval, others on a model’s existing knowledge; some display citations, others influence consideration without a measurable visit. Even apparently simple concepts need qualification. Being retrieved is not the same as being cited; being cited is not being mentioned; being mentioned is not being recommended. And being recommended does not guarantee a conversion.
Consequently, there cannot be one ideal profile for an AI Search expert. We need people who understand retrieval architecture, people who investigate system behavior, people who develop marketing methods suited to these changes, and people who explain complex discoveries without discarding their meaningful qualifications.
For this article, I have restricted the selection to professionals rooted in digital marketing. Academic researchers have made important contributions to generative information retrieval, but they are outside this particular selection.
The methodology: what earned someone a place?
I used six selection principles.
- Original contribution matters more than personal branding. I looked for experiments, datasets, technical investigations, tools, operational methodologies, original strategic frameworks, or unusually useful critical interpretations. A headline reading “GEO expert” is not evidence.
- The contribution must be traceable. Every profile links to publications, investigations, or other resources associated with that practitioner. Collaborative research must be credited as such: the person explaining a finding did not necessarily design the experiment or analyze the data.
- Different forms of originality deserve recognition. Building a knowledge-graph system is not the same as analyzing millions of AI responses. Neither is automatically more valuable than a strategic method that helps organizations make decisions under uncertainty.
- Evidence must be interpreted according to its limitations. Correlation is not causation. Simulated fan-out is not the same as observing Google’s internal process. Large datasets may still be unrepresentative, and results from one AI engine need not generalize to another.
- Commercial relationships are not disqualifying. Agencies and software vendors have access to valuable data and technical resources. Their studies deserve consideration, with their methods and commercial context open to scrutiny. I have also avoided giving one company’s publishing operation disproportionate representation.
- Expertise must be shared to be evaluated. I can already hear the objection: surely some extraordinary experts are less visible, perhaps because they never post on LinkedIn. Undoubtedly. But this is not a list of everyone who might know more; it is a list of practitioners whose contributions we can actually examine. Someone might be the next Tolkien, but if they never publish their Lord of the Rings, only their friends and parents can praise the masterpiece. A rigorous experiment read by fifty specialists matters more here than fifty viral posts. The criterion is not popularity. It is sharing useful knowledge with peers so it can be tested, challenged, and built upon.
Those principles lead to four categories. They are not hierarchical, and the names within them are not ranked. Several people could fit more than one group; I have assigned each to the contribution I regard as most distinctive.
| Category | People | Principal contribution |
|---|---|---|
| A — Technology and Methodology Builders | 8 | Tools, technical systems, and investigative methods |
| B — Experimental Researchers | 8 | Original evidence and experiments |
| C — Strategic Innovators | 7 | Marketing methodologies and strategic thinking |
| D — Divulgative Masters | 5 | Critical interpretation and communication |
A. Technology and Methodology Builders
These eight practitioners create tools, systems, or methodologies that help us investigate and operationalize AI Search. Some build semantic infrastructure; others reverse-engineer observable system behavior. They give the industry something tangible to examine, use, or extend, even when a tool remains an approximation rather than a proven model of a proprietary search engine.
1. Andrea Volpini — WordLift
Andrea Volpini belongs in any serious conversation about semantic search, knowledge graphs, and AI-driven discovery. As WordLift’s co-founder, he has spent years developing technology to make organizational knowledge more accessible to machines, well before the industry discovered GEO. His investigations into query fan-out and recursive language models over knowledge graphs connect earlier semantic-web thinking with contemporary AI retrieval.
Selected work: Query Fan-Out: A Data-Driven Approach to AI Search Visibility · Recursive Language Models and the Future of SEO.
2. Beatrice Gamba — WordLift
Beatrice Gamba represents a form of expertise the industry occasionally undervalues: turning complex semantic concepts into operational solutions. Her work covers knowledge graphs, ontologies, entity architecture, and the transformation of inconsistent commercial information into structured, machine-readable knowledge. Her Semantic Alchemy work shows how such relationships can become useful inside real business information systems. Her investigations into entity architecture and topical maps help connect established content strategy with emerging AI retrieval workflows.
Selected work: Semantic Alchemy: Turning Data into AI-Ready Knowledge · Why SEO Success Depends on Entity Architecture.
3. Dan Petrovic — DEJAN
Dan Petrovic approaches Search as a system to investigate, not a collection of tips to memorize. His fan-out research helps practitioners think beyond a user’s original prompt and consider the additional grounding searches that may contribute to a generated answer. His investigation of grounding snippets tackles another essential question: what content from a retrieved document actually becomes available to a generative system? It also makes clear why reverse-engineering observations must not be mistaken for full access to proprietary internals.
Selected work: Fanout Query Analysis · What Extraction Method Is Google Using to Build Grounding Snippets?.
4. Mike King — iPullRank
Mike King has been influential in connecting information retrieval concepts with practical SEO. His Relevance Engineering approach challenges the idea that optimization can be reduced to keywords and isolated page-level signals, emphasizing instead retrieval, relevance, and how a system connects an information need with evidence. His response to the “chunking” obsession is a useful reminder that passage retrieval does not prove every article should be cut into arbitrary blocks. His important contribution is a more rigorous way of investigating how Search works.
Selected work: There Is No “Accuracy” in AI Search Tracking – Only Precision · A Refutation of Misinformation About Chunking.
5. Suganthan Mohanadasan — Snippet Digital
Suganthan Mohanadasan investigates observable retrieval behavior instead of relying on what an AI assistant says it has done. His examinations of ChatGPT and Perplexity consider network activity, source selection, and the gap between information retrieved and citations displayed. The final answer reveals only part of the processing sequence; a citation does not necessarily tell us when a source was discovered or which alternatives were considered. Browser-visible traffic cannot expose the whole internal architecture, of course.
Selected work: How ChatGPT Actually Picks Sources · How Perplexity Actually Picks Sources.
6. Olivier de Segonzac — RESONEO
Olivier de Segonzac has conducted useful experiments examining ChatGPT’s observable search-tool behavior. His research explores the relationship between fan-out queries, discovery, caches, and the references ultimately shown to users. That addresses a persistent analytical mistake: believing the visible citations necessarily represent everything the system retrieved or evaluated. For practitioners building AI Search measurement systems, the separation of retrieval, processing, and citation presentation is foundational.
Selected work: Inside ChatGPT Search: Web Runs and Fan-Out Queries · Inside ChatGPT’s Retrieval Stack.
7. Jason Barnard — Kalicube
Jason Barnard has spent years developing operational methods for managing how search systems understand and represent entities, particularly brands. His work on Brand SERPs, Entity Homes, and the Kalicube Process predates the GEO boom. Its contemporary importance lies in the problem of machine-understandable identity: can a system distinguish a business from similarly named entities and reliably connect it to its attributes and subject areas? The terminology and performance claims of a proprietary framework should not be confused with independently confirmed mechanics of every AI engine.
Selected work: The Kalicube Process · Entity Home Methodology.
8. Mark Williams-Cook — Candour / AlsoAsked
Mark Williams-Cook combines technical curiosity, practical experience, and a welcome willingness to test fashionable assumptions. His QueryFan work explores how ChatGPT may expand an initial prompt into additional searches. Equally important is his critical approach to evidence. His cats.txt experiment challenged the causal reasoning behind claims that llms.txt was already a proven GEO optimization. It is a reminder that two things moving together does not establish that one caused the other.
Selected work: ChatGPT Is Secretly Googling Things · How cats.txt Exposed Weak llms.txt Evidence.
B. Experimental Researchers
These eight practitioners produce original observations, quantitative datasets, and experiments. Their work matters because the industry habitually turns observed correlations into proposed optimizations before it knows whether those relationships are causal. The strongest studies offer a better factual baseline, provided we remember the limits of their designs.
9. Metehan Yeşilyurt – PEEC AI
Metehan Yeşilyurt has distinguished himself through experiments with query fan-out, AI retrieval, and crawling behavior. His work includes simulated fan-out tools and Screaming Frog integrations that invite practitioners to test hypotheses against their own content. He has also investigated how fabricated statistics can spread across websites and AI answers. That latter work reveals risks of information contamination, although the method raises ethical concerns and should not be mistaken for a recommended tactic.
Selected work: Google AI Mode Optimization: Query Fan-Out · Query Fan-Out Analysis With Screaming Frog.
10. Nick Haigler — Seer Interactive
Nick Haigler has contributed empirical research into how AI systems represent and recommend brands. A study of more than 800,000 AI responses across nearly 2,000 brands investigated the relationship between third-party review profiles and AI visibility. His work on brand accuracy adds a second concern: appearing in an AI answer does not guarantee the information is correct. But correlations between review profiles and mentions do not establish that creating a particular profile causes greater visibility. That distinction is exactly why the findings are worth studying carefully.
Selected work: How Review Profiles Shape Brand Presence in AI Search · AI Brand Accuracy Study (coauthored with Bryan Gunawan).
11. Patrick Stox — SEO Consultant
Patrick Stox is one of the clearest examples of what large-scale quantitative analysis can contribute. His investigation of 55.8 million AI Overviews across 590 million searches provides a stronger basis for studying distribution patterns than manually collected screenshots. He has also investigated the challenges generative-engine referrals create for analytics. The caveat remains that even enormous datasets depend on collection methods, sampling, and dates; their size does not automatically guarantee representativeness.
Selected work: Insights From 55.8 Million AI Overviews · Generative Engines Are Breaking Web Analytics.
12. Kevin Indig — Growth Memo
Kevin Indig is especially interesting at the intersection of AI Search behavior, brand visibility, and commercial outcomes. His work asks what citation or mention metrics actually tell us about recommendations and user consideration. He also emphasizes the problems caused by prompt variation, context, reasoning processes, and model changes. His comparison of prompt monitoring with survey sampling, rather than deterministic rank tracking, deserves attention. His strength is making empirical observations meaningful to marketers while resisting the assumption that every count represents an outcome.
Selected work: AI Halftime Report: H1 2026 · Why Most Original Data Never Gets Cited.
13. Laurence O’Toole — Authoritas
Laurence O’Toole deserves recognition for bringing structured, longitudinal investigation into generative search early. Authoritas studied Search Generative Experience before AI Overviews became part of the routine Search landscape. Its research into SERP and AI Overview volatility distinguishes movement among source URLs from changes in the generated answer itself. A separate test with fabricated experts explored the difference between a chatbot acknowledging an allegedly authoritative person’s existence and recommending that person unprompted.
Selected work: SERP Organic and AI Overview Volatility Research · Can You Fake Expertise in AI Search?.
14. Glen Allsopp — Ahrefs
Glen Allsopp’s investigation into self-promotional “best” lists is directly relevant to the motivation for this article. By examining thousands of URLs appearing as sources for ChatGPT answers, he investigated how often commercially motivated listicles enter the information environment of AI recommendations. But a list appearing as a source does not prove an AI system accepts its rankings or endorses its author. Glen’s longstanding investigative approach makes him one of two Ahrefs practitioners I have included in this deliberately diversified selection.
Selected work: Do Self-Promotional Best Lists Boost ChatGPT Visibility? · Research Into AI Content Creation.
15. Joy Hawkins — Sterling Sky
Joy Hawkins supplies a perspective too often absent from GEO research: local businesses. Much of the industry’s attention goes to SaaS, ecommerce, and major brands, although AI-powered local interfaces can quickly change exposure for restaurants, healthcare practices, and location-dependent services. Her investigations, sometimes developed with Places Scout and Jepto, examine AI-influenced local results and their implications for calls and customer actions. The key insight is that stable traditional rankings do not necessarily mean stable commercial visibility when the interface itself is changing.
Selected work: The State of Local SEO in 2026 · How AI Is Changing Local Search Visibility.
16. Tomek Rudzki — Peec AI
Tomek Rudzki has produced quantitative investigations using large prompt and citation datasets. His study of 500,000 prompts examined when Google AI Overviews appear across different stages of the buyer journey, while other work explores ChatGPT fan-out patterns. Their limitation is that prompts in commercial visibility platforms may reflect the customers who select and track them, rather than the full universe of user needs. Within that boundary, the research remains useful for discovering patterns and developing more precise hypotheses.
Selected work: 500,000 Prompts and AI Overview Visibility · Patterns in ChatGPT Query Fan-Outs.
C. Strategic Innovators
A strategic framework becomes useful when it organizes decisions, identifies trade-offs, and connects technical possibilities with business objectives. The seven practitioners in this group contribute approaches to that problem. Creating a method and merely giving a popular name to someone else’s method are not the same thing.
17. Amanda King — FLOQ
Amanda King combines practical SEO strategy with a valuable critical perspective on AI Search measurement. Her presentation You Can’t Test Your Way Out of a Black Box challenges the idea that repeatedly querying opaque systems can produce deterministic ranking data. Her work on content consolidation also shows strategic thinking rooted in established search principles rather than fashionable terminology. I appreciate her interest in broader issues of transparency and AI training: marketing strategy cannot ignore the information environment in which these systems operate.
Selected work: You Can’t Test Your Way Out of a Black Box · What We Don’t Know About AI Training.
18. Wil Reynolds — Seer Interactive
Wil Reynolds has long argued that search marketers must understand people and business outcomes, not become obsessed with isolated metrics. Through Seer Interactive, he connects experimentation with trust, brand consideration, and commercial performance. His investigations into AI-assisted brand selection are valuable because they ask whether a generated answer changes the options a person seriously considers. They also illustrate the benefit of combining technical research with qualitative user research, including collaborators such as Andrea Haley. Visibility is interesting; changed behavior is the strategic question.
Selected work: ChatGPT Fan-Out Patterns and Brand (coauthored with Nick Aigler) · Can AI Change Which Brands Consumers Consider? (coauthored with Andrea Haley).
19. Aleyda Solís — Orainti
Aleyda Solís excels at transforming an emerging problem into a structured operational methodology. Her three-layer measurement framework separates AI presence, readiness, and business impact; a valuable distinction when dashboards casually combine citations, mentions, visibility estimates, traffic, and revenue proxies. It encourages marketers to connect observations with diagnosis, priorities, and outcomes. Her international SEO experience also matters because location, language, platform access, and audience context all complicate the meaning of an AI Search result.
Selected work: A Three-Layer AI Search Measurement Framework · AI Search Optimization Checklist.
20. Lily Ray — Amsive
Lily Ray has spent years examining search quality, reputation, trust, and organic visibility. Those questions become even more important when AI systems synthesize information and evaluate organizations for users. Her work connects established search-quality concerns with the new problem of brand presentation in generated answers. I particularly value her emphasis on distinguishing being cited as a source from being recommended as an option. Trust, useful content, and public reputation remain strategic concerns even when their means of exposure change.
Selected work: GEO, AEO and LLMO: Separating Fact From Fiction · It Works Until It Doesn’t: AI Content Strategies That Backfire.
21. Olaf Kopp — Aufgesang
Olaf Kopp has developed substantial work connecting information retrieval, entities, content strategy, and brand representation. His frameworks for LLM readability and brand-context optimization distinguish making information accessible from strengthening the associations between an organization and the subjects it wants to be known for. I would treat proposed causal mechanisms as strategic models to assess, rather than specifications of every AI engine. Nonetheless, Olaf offers a more structured alternative to the suggestion that GEO consists of adding short answers and FAQs to existing pages.
Selected work: Brand-Context Optimization · LLM Readability.
22. Dawn Anderson — Bertey
Dawn Anderson is among the strongest practitioners advocating a serious understanding of information retrieval. Her work connects established concepts – relevance, document selection, semantic matching, and retrieval evaluation – to contemporary generative Search. This perspective is important because many supposedly revolutionary GEO recommendations are reinterpretations of older ideas. I place her among strategic innovators because she relates technical knowledge to practical decisions, although her exceptional skill in explaining difficult material could equally place her among the Divulgative Masters.
Selected work: Debunking Generative Information Retrieval Misinformation · Why Information Retrieval Still Powers AI Search.
23. Crystal Carter — Wix
Crystal Carter brings the perspective of translating search practices into functionality and workflows on a large website platform. Her work connects technical SEO, structured data, and AI visibility with the realities of implementation across very different sites. Crystal instead examines which practices can be made repeatable and accessible for different organizations and levels of technical maturity. Her contribution is particularly strong in explaining where structured data serves established purposes and where AI-related performance claims still require evidence.
Selected work: 10 ways customer loyalty drives visibility in personalized search · How to optimize your website for the agentic web.
D. Divulgative Masters
I considered calling this group “Educators,” but the Italian word divulgazione better expresses what I mean: making specialized knowledge accessible without destroying its substance. A great divulgator does not merely simplify. They preserve meaningful distinctions, offer useful mental models, and recognize when the accurate explanation must include uncertainty. The people here also research, develop methods, and create ideas. Their classification recognizes the particular value of how they share them.
24. Rand Fishkin — SparkToro
Rand Fishkin has influenced how marketers think about Search since long before AI Search became a marketing category. His research on zero-click behavior is essential context for evaluating AI’s effects on publishers and brands. But his distinctive strength is explaining what changing discovery patterns mean for business. He clearly separates demand creation from demand capture and demonstrates why a platform can influence commercial decisions without generating traceable referrals.
Selected work: Less Than One Third of Google Searches Send a Click · Does Your Website Still Matter in the Zero-Click Era?.
25. Marie Haynes — Marie Haynes Consulting
Marie Haynes has a longstanding reputation for explaining complex search-quality developments to practitioners. Her analysis of information disclosed during Google’s antitrust proceedings connects technical questions about grounding and retrieval to implications for publishers and marketers. She also explores the emerging agentic web. Marie’s strength is explaining why a development matters without requiring readers to become information retrieval engineers. Her interpretations should, of course, be read alongside the primary disclosures from which they are drawn.
Selected work: FastSearch, MAGIT and Google’s AI · Why Google-Agent Represents a Major SEO Mindset Shift.
26. Jono Alderson — Independent Consultant
Jono Alderson consistently challenges the habit of treating each technological development as a pretext to invent another optimization checklist. His argument that “SEO vs. GEO” is the wrong question invites marketers to examine changing discovery, reputation, infrastructure, and consumer decisions rather than debate labels. His broader marketing manifesto criticizes channel-based thinking and visibility disconnected from commercial value. I put him here because his distinctive talent is helping practitioners interrogate the assumptions beneath their everyday work.
Selected work: SEO vs. GEO Is the Wrong Question · A Marketing Manifesto.
27. Mordy Oberstein — SE Ranking
Mordy Oberstein brings a brand-centered perspective to discussions often dominated by technical tactics. An AI system recommending a company is not simply retrieving a page containing particular keywords: it is presenting an interpretation of that organization. Mordy explains why inconsistent messaging and weak positioning may become problems in AI-mediated discovery, and why optimizing isolated pages cannot compensate for a poorly differentiated brand. His communication makes brand strategy relevant to practitioners who might otherwise think GEO is primarily a technical discipline.
Selected work: How to Better Measure LLM Visibility · LLM Visibility Starts With Better Internal Communication.
28. Andy Crestodina — Orbit Media
Andy Crestodina excels at connecting research with decisions. His investigation of AI-referred traffic examined nearly 29 million sessions across 97 B2B and lead-generation websites, considering traffic behavior and conversion characteristics while documenting limits such as inconsistent conversion tracking. His research into how marketers are responding to AI Search adds another useful perspective. Andy explains analytical findings without losing sight of what they can—or cannot—tell a business.
Selected work: AI Traffic Conversion Research · What 1,000+ Marketers Are Doing About AI Search.
Honorable mentions: Ten more practitioners worth following
The selection could easily have included these ten practitioners. Their work is worth reading alongside the main profiles; the distinction here is editorial space, not a verdict on their expertise.
Dixon Jones — InLinks / Waikay. A pioneer of entity-led SEO whose tools and research connect semantic relationships with AI-era brand recognition. What you should read: A Cynical Read of Google’s AI Optimisation Guide.
Alex Moss — Yoast. Brings practical technical SEO and structured-data expertise to the questions of AI visibility and agentic commerce. What you should read: Why Data Integrity Is The New Technical SEO: From Crawling To Trust.
Myriam Jessier — PRAGM. Combines technical branding, multimodal search, and a rigorous distinction between being cited and being recommended. What you should read: Brand Depth Determines What AI Systems Recommend.
Aimee Jurenka — RicketyRoo. Documented how a focused topic-entity restructuring coincided with improved AI visibility for a local business. What you should read: The Role of Informational Content in the Age of LLMs.
Lazarina Stoy — MLforSEO. Connects NLP, machine-learning workflows, and personalization to the changing mechanics of search retrieval. What you should read: How AI Search Personalizes Fan-Out Queries Using Memory and Context.
Chris Green — Torque Partnership. Tests the practical limits of AI-assisted SEO, reminding us that what an assistant can render is not necessarily what its search tools retrieve. What you should read: If It’s Important (to AI), Put It in the Raw HTML.
Duane Forrester — UnboundAnswers. Brings former search-engine insider experience to a broader view of how AI changes discovery, consideration, and buying behavior. What you should read: AI Changed Your Buyer Faster Than Your Business Can Adapt.
Cindy Krum — MobileMoxie. Long ahead of the curve on mobile, multimodal, and journey-based search, with a critical eye on concentration and personalization. What you should read: AI Search and the Shift Toward Inauthenticity and Commercial Interests.
Gus Pelogia — Indeed. Grounds AI Search advice in product-led SEO, realistic measurement, and a repeatable way to investigate brand visibility. What you should read: How to Reverse-Engineer LLM Brand Visibility.
David Konitzny — Peec AI. Uses direct technical observation to examine retrieval pipelines and the limits of what AI Search visibility metrics reveal. What you should read: How ChatGPT Deep Research Reads Your Site: What the Logs Reveal.
One final disclosure: Why I’m not on this list
You may have noticed one conspicuous absence. Mine. Yes, I know: an inexplicable editorial decision.
I’ve worked in SEO since 2004, developed an AI Search Optimization Framework, and spent a slightly unreasonable amount of time investigating semantics, entities, international Search, and retrieval… the sort of subjects guaranteed to make you extremely popular at dinner parties. I could make a convincing case for myself. In fact, I would probably be the ideal person to write it. Which is precisely the problem this article started by criticizing.
So I’ll leave the imaginary medal in its box. Anyone who wants to judge the work can start with my AI Search Optimization Framework and Transmedia Storytelling in the Age of AI Search on this same blog, plus International AI Search SEO and my Guide to Images & Visual Search for Advanced Web Ranking.
Not a bad little collection, if I may say so myself. But I’ve already said too much about myself in an article explaining why people shouldn’t nominate themselves. Back to the interesting part.
What these 28 practitioners reveal about the future of AI Search
Looking at these people collectively is more revealing than looking at their individual contributions. We do not discover one unified school of GEO, nor a checklist of tactics everyone endorses. Instead, several perspectives on how AI is changing Search emerge. They overlap, reinforce one another, and sometimes expose useful disagreements.
1. Search is increasingly interpreting needs, not just matching queries to documents
Mike King, Dan Petrovic, Suganthan Mohanadasan, Olivier de Segonzac, and Metehan Yeşilyurt direct our attention to what happens between a user’s initial request and the generated answer. A system may reformulate the question, fan out into additional searches, retrieve information from varied sources, evaluate passages, and combine evidence.
The deeper issue concerns the unit of relevance. A query is increasingly one element in a broader context that may include earlier interactions, implicit constraints, geography, and the system’s interpretation of intent. Two superficially identical starting queries might therefore lead to different information journeys. It is also why personalization and contextual Search deserve serious investigation.
2. Knowledge architecture matters, but a brand’s information environment extends beyond its site
Andrea Volpini, Beatrice Gamba, Jason Barnard, Olaf Kopp, and Dawn Anderson approach the problem differently, but all focus on information representation and relationships.
Structured data, ontologies, architecture, and clear entity identification can help organize that knowledge. They do not automatically guarantee that a specific AI system uses it. And the relevant information extends beyond the organization’s domain: specialist publications, associations, independent journalism, professional communities, reviews, and other third-party sources help establish a wider informational context.
This suggests closer collaboration among technical SEO, content strategy, digital PR, and brand development, not because those disciplines become interchangeable, but because the evidence AI systems might use is distributed among them.
3. Brand visibility is not the same as brand preference
Kevin Indig, Wil Reynolds, Lily Ray, Glen Allsopp, and Laurence O’Toole approach a fundamental question from different angles: what does it actually mean for a brand to appear in AI Search? A company may be retrieved but not cited, cited but not named, named but not recommended. Even an explicit recommendation may have very different effects according to trust, context, and the user’s previous familiarity with the organization.
A source appearing in an answer is not proof that the AI endorses that source’s ordering or that the end user accepts its claims. Marketers need to separate source visibility, entity recognition, recommendation visibility, and commercial influence. The more important strategic question may be how the answer changes someone’s evaluation of the options available.
4. AI Search measurement is becoming probabilistic
Amanda King, Aleyda Solís, Patrick Stox, Kevin Indig, and Tomek Rudzki, among others, challenge the assumption that AI Search can be measured like conventional rank tracking. Generated answers add variability through prompt wording, conversation history, personalization, retrieval decisions, and model behavior.
Repeating a prompt does not necessarily reproduce the same information need under the same conditions. And a visibility score drawn from a small, commercially selected prompt set is not an unbiased estimate of audience demand. Measurement is possible, but it requires transparent sampling, appropriate uncertainty, and a distinction between observed facts and estimated influence.
5. Zero-click discovery changes the economics of visibility
Rand Fishkin, Andy Crestodina, Joy Hawkins, and Wil Reynolds reveal a problem that pure retrieval analysis cannot resolve. AI interfaces may satisfy informational needs without sending visitors to a publisher, but the consequences depend strongly on the business model. An AI answer can influence brand perception or purchase consideration without producing a measurable visit. For a publisher dependent on advertising, the same absence of clicks can represent a significant commercial loss.
Local Search adds another variation: changes in the interface may change which businesses appear and which actions users can take, even if conventional rankings look stable.
6. From answering questions to carrying out tasks
A further pattern becomes apparent when we consider semantic infrastructure, platform implementation, and strategic thinking about agentic systems. Much GEO commentary assumes that the final product is an answer. But what changes when a system’s objective is to complete a task?
Recommending accommodation is different from checking availability, comparing cancellation policies, choosing an appropriate rate, and initiating a reservation. The second scenario depends on reliable inventory, structured information, interfaces, business rules, and confirmation mechanisms. Discoverability may increasingly involve not just presenting knowledge but participating in an operational workflow.
7. The ability to challenge assumptions may be our most valuable skill
Mark Williams-Cook, Dawn Anderson, Amanda King, Marie Haynes, and Jono Alderson remind us that we are studying systems whose mechanics remain partially hidden and change frequently.
A discipline does not mature simply by producing more hypotheses. It advances when people design better tests, reject explanations unsupported by results, and distinguish what is known from what is merely plausible. This is also why exceptional divulgation matters. Explaining the limits of a discovery can be just as important as announcing the discovery itself.
So, is there a common vision?
I do not think these 28 practitioners share one vision, and that is precisely what makes their collective work interesting. Some concentrate on retrieval architecture; others on knowledge representation, brand recommendations, measurement, or the operational demands of agentic systems.
What connects the strongest contributions is a recognition that AI Search cannot be adequately understood as only traditional SEO with a different interface and a new acronym.
It also explains why the discipline needs all four forms of expertise. Builders create systems and investigative methods. Researchers produce evidence. Strategists connect evidence with decisions. Divulgators help peers understand, question, and extend what has been learned.
That brings me back to this article’s purpose: recognizing those who make expertise available for others to test, challenge, and build upon.
Expertise deserves respect. Shared expertise advances the discipline.
Editorial note: This selection reflects publicly documented work available as of October 2026. It is a curated professional assessment, not a definitive ranking. Category assignments are subjective, and inclusion does not endorse every conclusion made by the featured practitioners. Research should be read in light of its methods and limitations, and collaborative contributions deserve proper attribution.
