LinkedIn voices on AI search and GEO to follow in 2026

AI search changes every month, and most of what is written about it is opinion. The fastest way to keep up is to follow a few people who test things and publish the results. These are eight people worth following on LinkedIn for AI search, GEO and AI Overviews, with what each one is known for and one thing you can learn from them.

Who should you follow on LinkedIn for AI search and GEO?

The list is not ranked. It runs from search data to retrieval, measurement and distribution, so you can read it as a path. Each person was checked on 3 November 2026.

1. Lily Ray: watch the data when Google changes something

Lily Ray is VP of SEO and AI Search at Amsive and founded the consultancy Algorythmic. On LinkedIn she posts about Google updates, AI Overviews and AI Mode, often with charts from real sites. In September 2026 she posted a site's clicks from France around the day AI Overviews rolled out there: a drop of 87% in clicks month over month, with a matching rise in AI Overview impressions. Her recurring point is that strong SEO fundamentals still power AI search.

What to learn: when Google launches something, open your own Search Console and compare clicks and impressions before and after. One site is not a study, but your own data beats anyone's general claim.

2. Mike King: understand how retrieval works

Mike King is the founder and CEO of iPullRank and coined "Relevance Engineering". His team publishes The AI Search Manual, a long guide to how AI search finds, chooses and generates answers. In September 2026 he wrote on LinkedIn that someone at an EMARKETER conference asked him which objective data source to check AI visibility against. His answer: there isn't one. AI search tracking has no accuracy, only precision, so you have to scrutinize each tool's method.

What to learn: before you pay for an AI visibility tool, ask which prompts it runs, how often it repeats each one, and on which engines. A score without a method is a guess.

3. Aleyda Solís: turn AI search into a roadmap

Aleyda Solís is an SEO and AI search optimization consultant, runs the boutique consultancy Orainti and writes the SEOFOMO newsletter. Her Crawling Mondays series has a run of episodes on AI search optimization, each with practitioners. In September 2026 she shared an episode on the most common AI search optimization spam, myths and risky tactics, with Dawn Anderson and Gerry White among the guests.

What to learn: start with how your own audience uses AI search, then pick the topics, then measure. And keep a list of the tactics experts call myths, so you can filter the next "secret GEO trick" you see.

4. Kevin Indig: test the same question everywhere

Kevin Indig writes the Growth Memo newsletter, where he publishes research on SEO and AI search. With Eric van Buskirk he ran the first UX study of Google's AI Overviews (May 2025): 70 users across 8 search tasks. In October 2026 he warned on LinkedIn against using ChatGPT as a proxy for all AI search: ChatGPT is still ahead, but Gemini gained adoption and Claude started climbing faster in early 2026. That post is labeled as a brand partnership on LinkedIn.

What to learn: take a question a buyer asked you on a sales call. Ask it in ChatGPT, Gemini and Claude, see who gets recommended, and read the sources behind each answer. You may find a competitor that barely shows up in one engine and leads in another.

5. Rand Fishkin: stay skeptical

Rand Fishkin is co-founder and CEO of SparkToro, an audience research company, and writes about zero-click marketing. In September 2026 he listed five reasons he is skeptical of the AI tracking industry: AI tools give different answers to the same question every time, and there is no source of truth for how often people ask a given prompt. On SparkToro's blog (October 2026) he argues that AI should be funded and measured as a brand channel, "a digital billboard", not as a traffic source.

What to learn: one AI answer is one sample. Look at trends over many runs, and judge AI visibility by what it does for your brand, not by referral clicks alone.

6. Britney Muller: learn how the models work

Britney Muller founded Orange Labs, which teaches people and companies to use AI with confidence. Her background is machine learning, technical SEO and research at Moz and Hugging Face. Her LinkedIn posts explain how LLMs work in plain language, and she is open about the limits: in a recent post she wrote that she spends a surprising amount of time telling clients they don't need AI.

What to learn: an LLM predicts text; it is not a search index. Once you understand that, you stop expecting one prompt to show your "ranking", and you see why being mentioned often and consistently matters.

7. Chris Long: read the fan-out queries

Chris Long is co-founder of Nectiv, an SEO and GEO agency for B2B and technology companies. At brightonSEO he presented an analysis of more than 28,000 fan-out queries from ChatGPT and Gemini: the searches an AI engine runs behind the scenes to answer one prompt. On LinkedIn he shares new AI search studies quickly, often with a short take on what they mean for B2B sites.

What to learn: a single prompt becomes several searches. Ask which sub-questions an engine would need answered to recommend you, and make sure your site and LinkedIn answer each one directly. How that works on a page is in the guide on how to get ChatGPT to cite you.

8. Ross Simmonds: distribute, don't just publish

Ross Simmonds is CEO of Foundation, which describes itself as an AI visibility agency for B2B tech and SaaS, and of Distribution.ai. He wrote "Create Once, Distribute Forever". On LinkedIn he argues that distribution matters even more now that AI splits search across many platforms, and in October 2026 he shared a study Foundation ran with AirOps on how B2B SaaS brands show up in AI answers.

What to learn: AI engines cite what others say about you. Muck Rack found that 84% of the links ChatGPT, Claude and Gemini cited came from earned media (May 2026). So one strong piece of research, spread to LinkedIn, Reddit, YouTube and the trade press, does more than ten posts on your own blog.

How do you choose who to follow?

Many of the "top AI SEO influencer" lists in Google on 3 November 2026 came from agencies and tool vendors, ranked by follower counts or a proprietary score, and some included their own founders. These are the criteria I used instead, so you can apply them to anyone:

1. They post about AI search on LinkedIn now. A recent post on AI search, GEO, AI Overviews or LLM visibility, not just an old reputation. One person I considered was left out because the last LinkedIn post I could find was seven months old.
2. They show their work. Their own data, a test, a study or a named method, not only predictions.
3. They exist outside LinkedIn. A company, a study, a newsletter or talks you can check.
4. They say what they sell. Most of them run an agency or a company. That is fine when it is clear, and a post labeled "brand partnership" is a signal to read the claim twice.
5. They are allowed to disagree. A good list has skeptics and builders. Here, Rand Fishkin and Mike King question AI visibility tracking, while others build methods around it.

What I left out on purpose: follower counts (they change daily and measure reach, not rigor), a ranking (there is no fair metric for "best"), and anyone who promises guaranteed citations in ChatGPT. No engine sells that.

How do you learn from them without copying?

Following experts only helps if you turn what they post into your own work. Copying their takes makes you one more echo, and AI engines have no reason to cite an echo. This is the routine that works:

1. Save the method, not the sentence. When a post describes a test, write down the steps, not the wording.
2. Run it on your own data. Kevin Indig's three-engine test with your sales-call question, or Lily Ray's before-and-after look at Search Console. Your result is new information; their result is not.
3. Write what you found, with your experience in it. What surprised you, what didn't match their result, what you will change. That is the part only you can write.
4. Credit them by name. Tag or mention the person whose method you used. It is honest, and they often join the conversation.
5. Put the full version on your site and a shorter one on LinkedIn. A different, shorter LinkedIn version that links to the full answer in its first line. We do this on our own site; the data is in our SEO and GEO case study.

What do these experts agree on?

For all their arguments, three ideas come up again and again in their posts:

• SEO fundamentals still count. Pages that rank and answer clearly are the pages AI engines can find. The full picture of how GEO and SEO fit together is in the GEO guide.
• Measurement is fuzzy. AI answers vary from run to run, so treat any visibility score as a trend, not a fact.
• You need to be mentioned, not only published. Engines lean on what other sites and people say about you, which is why personal LinkedIn profiles and earned media matter.

The rest of the series, from AI Mode to AI Overviews, is on the AI visibility hub. If you are a CEO or founder and want to be one of the voices AI engines quote in your own field, that is what CEO personal branding builds: a LinkedIn body of knowledge in your voice, around three or four topics.