AI Can't Agree Who the Best AI SEO Teacher Is. We Ran 90 Queries to Find Out Why.
90 queries across ChatGPT, Gemini, and Perplexity asking for the best AI SEO creators. Three different #1 winners came back. And Julian Goldie — who wins the 'tactical' intent outright — never shows up on ChatGPT at all.
By AIAttention Research
Last week we asked ChatGPT, Gemini, and Perplexity the same question six different ways. Ninety queries total. We were expecting one clear answer — who's the best AI SEO creator to learn from? — and a tidy leaderboard.
Three different #1s came back. And one of them never showed up on ChatGPT at all.
Here's what we found.
The Setup
We run AIAttention, a tool that tracks brand visibility inside AI answers. A natural question for it: who dominates AI's mental model of "AI SEO experts"?
We built six phrasings of the same underlying question:
- Best — "…best for learning AI SEO in 2026?"
- Popular — "…most widely followed by marketers?"
- Beginner — "…most beginner-friendly?"
- Advanced — "…best for advanced practitioners?"
- Tactical — "…most tactical with real case studies?"
- Trusted — "…most trusted by experienced SEO teams?"
Each one ended with Exclude agencies, software tools, and companies. so the responses would return people, not brands.
Each prompt ran through ChatGPT (web), Gemini (web), and Perplexity (web) five times. 90 queries total. A GPT-4o-mini extractor pulled creator names from each raw response in order, and each mention got a position-weighted score: 1.00 for first, 0.75 for second, 0.56 for third.
Three Different #1s Across Six Intents
| Intent | #1 creator |
|---|---|
| Best | Matt Diggity |
| Popular | Matt Diggity |
| Beginner | Nathan Gotch |
| Advanced | Matt Diggity |
| Tactical | Julian Goldie |
| Trusted | Nathan Gotch |
Matt Diggity wins three intents. Nathan Gotch wins two. Julian Goldie wins one.
But the interesting number is the spread within a creator. Matt Diggity's weighted score in Advanced is 10.82. In Trusted it's 1.88. Same creator, two phrasings of the same underlying question, a 5.8x swing in how much AI recommends him.
If you asked only "Which creators are best for learning AI SEO?" you'd meet Matt Diggity and assume he owns the category. You'd never meet Lily Ray. You'd never meet Koray Tuğberk GÜBÜR.
ChatGPT and Gemini Recommend Different Worlds
This is the part we didn't expect.
ChatGPT's consistent top creators: Brian Dean (Backlinko), Matt Diggity, Neil Patel, Ryan Stewart, Miles Beckler.
Gemini's consistent top creators: Nathan Gotch, Matt Diggity, Aleyda Solis, Koray Tuğberk GÜBÜR, Julian Goldie.
Only Matt Diggity reliably appears on both lists. The rest barely overlap.
The pattern is legible:
- ChatGPT tilts toward US-establishment SEO — Backlinko, Neil Patel, Ryan Stewart fit a specific mold.
- Gemini tilts toward international practitioners — Koray GÜBÜR (Turkey), Aleyda Solis (Spain), Julian Goldie (UK/Australia).
You can test this yourself. Ask ChatGPT: "which individual YouTube creators give the most tactical AI SEO advice with real case studies?" You'll probably get an apology and a non-answer. Ask Gemini the same thing: you'll get Julian Goldie, Matt Diggity, Koray GÜBÜR, Jacky Chou — four named creators, in rank order.
If you're a brand, tracking one model isn't tracking "AI visibility." It's tracking one model's worldview.
Lily Ray Appears Only in "Trusted"
Lily Ray showed up 12 times out of 90 queries. Every single mention was in the Trusted intent. She doesn't crack top-8 in Best, Popular, Beginner, or Tactical.
She's not a high-volume YouTube uploader. Her footprint is X posts, LinkedIn writing, Amsive's consulting, conference talks. AI has clearly decided: when someone asks about trusted experts, authority signals outweigh YouTube activity.
Same pattern shows up for:
- Kevin Indig (Trusted only)
- Marie Haynes (Advanced + Trusted)
- Rand Fishkin, Mike King (iPullRank), Cindy Krum, Barry Schwartz — all Trusted only
The Trusted intent pulls a different tier: respected voices whose influence doesn't depend on an upload schedule. Every other intent rewards active YouTubers.
Julian Goldie Is Invisible on ChatGPT
Across all 30 ChatGPT queries — every prompt, every run — Julian Goldie is mentioned zero times.
On Gemini he's top-3 in Popular, Tactical, Beginner, Advanced. On Perplexity he's top-5 in Popular and Tactical.
This is the starkest per-creator model split in the dataset. If Julian's audience is growing, it's growing inside Gemini and Perplexity — and ChatGPT users don't know he exists.
The inverse: Ryan Stewart is strong on ChatGPT (top-3 in Popular and Advanced) and almost absent on Gemini.
These aren't edge cases. They're structural splits in AI's mental model of the creator landscape.
The Top 10 Overall
| Rank | Creator | Weighted | Intents won |
|---|---|---|---|
| 1 | Matt Diggity | 35.90 | Best, Popular, Advanced |
| 2 | Nathan Gotch | 29.21 | Beginner, Trusted |
| 3 | Julian Goldie | 18.09 | Tactical |
| 4 | Brian Dean | 17.33 | — |
| 5 | Neil Patel | 14.78 | — |
| 6 | Backlinko | 10.74 | — (same as Brian Dean) |
| 7 | Koray Tuğberk GÜBÜR | 10.60 | — |
| 8 | Aleyda Solis | 7.47 | — |
| 9 | Lily Ray | 6.04 | — |
| 10 | Marie Haynes | 6.32 | — |
"Backlinko" and "Brian Dean" are the same creator — the interpret engine kept them as separate canonical names. Merged, Brian Dean moves to #3.
What This Means If You're Competing for AI Attention
AI chat is replacing Google for "who should I learn from?" questions. That means creators and brands now compete for mention inside an AI answer, not for a SERP position. The rules of that competition are different.
Three from this experiment:
There is no single score. Aggregate visibility hides structural patterns. Strong in Popular + invisible in Trusted is a different problem than strong in Tactical + weak in Best.
The model matters more than the phrasing. ChatGPT and Gemini don't know the world slightly differently. They know different people.
Authority and activity are measured separately. The Trusted intent pulls a wholly different tier — people whose influence is written, not uploaded. Different competitive landscape, different signals to optimize for.
What's Next
Phase 2 runs in May. Same 6 intents. Adds YouTube metrics — subscriber count, recent-video engagement, upload cadence — for the full tracked pool.
The question: does AI track subscriber count? Recent traction? Or is it looking at something else entirely?
Phase 1 hints at the answer. Lily Ray wins Trusted without a large YouTube presence. Julian Goldie wins Tactical with a Gemini-only audience. Matt Diggity wins Advanced the traditional way — big channel, sustained output, high signal across models.
Phase 2 writeup publishes in about a month.
Want to see where your brand shows up across ChatGPT, Gemini, Perplexity, and Qwen — across the intents your customers actually use? Track it at aiattention.ai.
Questions or intent phrasings you'd want tested next? Email us at hello@aiattention.ai.
Start measuring your AI visibility today. Get Started Free →