This repository is a structured collection of primary-source research on AI-powered SEO content production, drawn directly from practitioners who actively implement AI search strategies rather than commentators writing about them secondhand. It includes LinkedIn posts, YouTube transcripts, expert-level summaries, and a curated source directory, all organized to surface common frameworks, practical workflows, and emerging trends that can inform a future AI-powered SEO content playbook.
AI-powered SEO is reshaping how content is created, optimized, discovered, and cited by systems like ChatGPT, Gemini, Claude, and Google's AI Overviews. Understanding how experienced practitioners approach AI search visibility, Generative Engine Optimization (GEO), Answer Engine Optimization (AEO), AI citation strategy, semantic SEO, and AI-assisted content production provides a grounded basis for building scalable, effective content strategies rather than relying on secondhand trend pieces.
Experts were selected for demonstrated, hands-on engagement with AI search rather than general popularity. The core filter was simple: does this person actually build things, run experiments, and publish results, or do they primarily comment on what others are doing? Every expert listed below was verified as actively publishing in 2025 and 2026, and each brings a distinct, non-overlapping angle to AI-powered SEO content production that the others do not cover.
| Expert | Focus Area |
|---|---|
| Dan Petrovic | Research-driven AI search mechanics and forward-looking SEO trends |
| Bernard Huang | GEO, AEO, and AI citation optimization |
| Britney Muller | LLM behavior and information retrieval research |
| Eric Siu | Applied AI workflows across content and marketing operations |
| Kevin Indig | Data-driven research on AI search and growth strategy |
| Koray Tugberk Gübür | Semantic SEO and topical authority |
| Lily Ray | E-E-A-T, AI search, and content quality |
| Michał Suski | AI search visibility and content optimization |
| Nathan Gotch | Practical AI SEO frameworks and experiments |
| Olga Zarr | AI Overviews, AI visibility, and technical SEO |
- Dan Petrovic was chosen because he is producing the most technically rigorous original research of anyone in this field. He coined "Selection Rate Optimization," reverse-engineered Google's shopping classifier, ran a 14,000-prompt experiment measuring brand recall patterns across Gemma and Gemini, and built a free public toolkit of over twenty AI search tools covering everything from citation mining to token-level entropy analysis. He has held two Google awards for discovering internal APIs and algorithms. Most of the industry cites work like his secondhand. He was chosen to give this research a genuine technical foundation rather than a practitioner-opinion layer.
- Bernard Huang was chosen because he spent nine years building Clearscope, the tool a large portion of the content and SEO industry actually uses for AI-era content optimization, before stepping away in May 2026 to build autonomous AI businesses. His perspective on AEO and AI citations comes from watching real performance data across thousands of customer accounts, not from theorizing. He also introduced a concrete, repeatable workflow for AI citation strategy that is more actionable than most of what circulates in the space: identify your core commercial prompts, find what the AI actually searches for when answering those prompts, then target the high-authority sources it pulls from rather than trying to influence the model directly.
- Britney Muller was chosen because she has genuine machine learning depth that almost no one else in the SEO space can match. She led the launch of BLOOM, one of the largest open-source multilingual language models, at Hugging Face, and spent four years as Senior SEO Scientist at Moz before that. This gives her a direct understanding of how LLMs actually work at a technical level, which she uses to push back clearly on misconceptions the industry keeps recycling, particularly the idea that LLMs "rank" content the way Google does, or that you can reverse-engineer AI visibility the way you can reverse-engineer an algorithm.
- Eric Siu was chosen because he runs AI-assisted content production at real agency scale, not as a demo or experiment. He documented generating over 130 production-grade SEO pages in 24 hours using AI workflows, built six named specialist AI sub-agents handling SEO, CRO, paid media, and ad creative at Single Grain, and shared specific results which includes social content agent generating posts with 300,000 to 500,000 views, an analytics agent finding $500,000 in cost savings within three days of deployment. He is one of the few voices in this space where the AI workflow claims are backed by a named client business operating at enterprise scale.
- Kevin Indig was chosen for two reasons. First, his operator background is to lead SEO and growth at Shopify, G2, and Atlassian, which means his frameworks for AI search strategy come from having to make real decisions with real budgets and real stakeholders, not from consulting or media. Second, his methodological rigor around prompt tracking and AI visibility measurement is unusually specific: he publishes actual sample-size recommendations, per-platform scoring guidance, persona segmentation logic, and confidence intervals, rather than the vague "track your brand mentions" advice most others offer.
- Koray Tugberk Gübür was chosen as the originator of the Topical Authority framework, which underpins a significant portion of how the industry now thinks about semantic SEO and content coverage at scale. He has published over twelve SEO success stories, ten case studies, and two formal A/B tests with real before and after data, and has worked with more than 200 companies across Europe, Africa, and North America. He brings a content systems angle on how to structure a site's entire information architecture to communicate authority to both search engines and AI models, that complements the citation-level and technical-level perspectives of the other experts.
- Lily Ray was chosen specifically because she is the strongest documented voice on where AI content production goes wrong, not just where it works. She has tracked dozens of real, named cases where brands scaled AI content or GEO tactics, gained visibility quickly, and then saw both their Google rankings and their ChatGPT citations collapse together with visibility data to support each case. She was also among the first to publicly document the January 2026 Google update penalizing self-promotional listicle content at scale, and is actively researching how ChatGPT and Claude are reducing spam and manipulation in their own citation behavior. A research set that only captures the upside of AI content production without this perspective would be incomplete.
- Michał Suski was chosen because he co-founded Surfer and has spent over eight years building data-driven content optimization tooling used by more than 150,000 creators. His live "AI Search Visibility Roast" sessions, where he analyzes real brands on screen in real time, are among the most hands-on, evidence-based content in this space showing exactly where specific brands (IKEA, Nike, Slack) are winning or losing AI citations and why, with the data visible throughout. He also introduced "information retrieval cost" as a distinct optimization concept, backed by internal Surfer data showing that consistently-cited sources have meaningfully higher fact coverage than non-cited ones, independent of domain authority.
- Nathan Gotch was chosen because he builds software specifically for the problem this research addresses. Rankability, his platform launched in 2024, is designed to give agencies predictable visibility across Google, ChatGPT, and AI search from a single workflow. His content reflects patterns across many client accounts simultaneously, which makes his findings more generalizable than a single-agency perspective. His HubSpot case study shows that 89% of AI citation sources for the "best CRMs" query mentioned HubSpot. Despite HubSpot's own domain not appearing as a direct citation, with 72% of those mentions being unlinked, which is one of the most specific and reproducible data points in this entire research set.
- Olga Zarr was chosen to provide a grounding, reality-check perspective alongside the more forward-leaning voices. With over fourteen years of hands-on SEO consulting, she consistently pulls the conversation back to verifiable data: Google handling roughly 373 times more daily queries than ChatGPT, AI Overviews appearing in over 50% of searches while cutting click-through rates by 18 to 64% depending on query type, and AI crawler behavior differing in specific, testable ways from traditional Googlebot. Her SEO Audit Mastery approach, and her newsletter reaching over 10,000 subscribers, means her content is consistently grounded in what practitioners actually encounter working inside real accounts rather than what makes for a good conference talk.
For each expert, the following material was gathered and organized:
LinkedIn content : Recent posts covering AI search discussions, GEO and AEO frameworks, AI citation strategies, and real-world case studies, stored under research/linkedin-posts/.
YouTube content : Interviews, conference presentations, tutorials, and AI search strategy discussions, stored under research/youtube-transcripts/.
Source directory : A centralized expert reference (research/sources.md) containing each expert's name, profile links, content sources, and a brief annotation explaining their specific relevance to this project.
Insights : Structured summaries of both LinkedIn posts and YouTube transcripts, covering expert name, title, date published, duration, URL, summary, and key insights for each piece of content, stored under research/insights/.
Supplementary materials : Two synthesized documents built from the collected research, stored under research/other/:
- content-production-framework.md consolidates the most actionable frameworks, workflows, and tactical recommendations from across all ten experts into a single structured reference. It covers how practitioners currently approach AI-assisted content creation, optimization for AI citations, prompt-based workflows, and quality control which is distilled from the raw LinkedIn and YouTube material into a usable framework layer.
- ai-seo-playbook-outline.md is a structured outline for building a full AI-powered SEO content playbook, organized into strategic phases. It maps the key themes, frameworks, and expert insights from this research into a logical sequence covering AI search fundamentals, content production systems, citation optimization, performance tracking, and iterative improvement which is designed to be used as the structural backbone of an eventual full playbook.
AI-Powered-SEO-Content-Production/
│
├── README.md
│
├── research/
│ ├── sources.md # Expert directory with links, roles, and annotations
│ │
│ ├── linkedin-posts/ # One folder per expert, one .md file per post
│ │ ├── dan-petrovic/
│ │ ├── bernard-huang/
│ │ ├── britney-muller/
│ │ ├── eric-siu/
│ │ ├── kevin-indig/
│ │ ├── koray-tugberk-gubur/
│ │ ├── lily-ray/
│ │ ├── michal-suski/
│ │ ├── nathan-gotch/
│ │ └── olga-zarr/
│ │
│ ├── youtube-transcripts/ # One folder per expert, one .md file per video
│ │ ├── dan-petrovic/
│ │ ├── bernard-huang/
│ │ ├── britney-muller/
│ │ ├── eric-siu/
│ │ ├── kevin-indig/
│ │ ├── koray-tugberk-gubur/
│ │ ├── lily-ray/
│ │ ├── michal-suski/
│ │ ├── nathan-gotch/
│ │ └── olga-zarr/
│ │
│ ├── insights/ # Structured summaries of all collected content
│ │ ├── linkedin-posts-summary.md # Expert, title, date, URL, summary, key insights
│ │ └── youtube-transcripts-summary.md # Expert, title, date, duration, URL, summary, key insights
│ │
│ └── other/ # Synthesized outputs built from the research
│ ├── content-production-framework.md # Consolidated frameworks and workflows across all experts
│ └── ai-seo-playbook-outline.md # Structured outline for a full AI-powered SEO playbook
│
└── scripts/
├── scrape_linkedin.py # LinkedIn post extraction script
└── fetch_youtube_transcripts.py # YouTube transcript collection script
Development environment : Cursor IDE with Cursor Agent for AI-assisted coding and debugging, Python 3.10, Git, GitHub
Python libraries : youtube-transcript-api, requests, beautifulsoup4, lxml, pyyaml, yt-dlp
AI tools : Cursor Agent, Claude Code / Codex, ChatGPT
Data sources : LinkedIn, YouTube
Step 01: Repository Setup : Created the GitHub repository (AI-Powered-SEO-Content-Production), cloned it locally, and opened it in Cursor IDE. Established the initial project structure with research/ and scripts/ directories before any content collection began.
Step 02: Expert Research and Source Documentation : Identified ten leading practitioners across AI SEO, GEO, AEO, and AI search, then gathered their LinkedIn profiles, post URLs, and relevant YouTube video links. Documented all expert profiles, source links, and metadata in research/sources.md. Committed and pushed the initial repository structure to GitHub.
Step 03: LinkedIn Post Collection :
Built scripts/scrape_linkedin.py using Cursor Agent and configured it with the selected LinkedIn post URLs. The script was iteratively refined to support multiple extraction methods including JSON-LD parsing, SocialMediaPosting structured data extraction, video transcript extraction, HTML fallback parsing, and Open Graph metadata that helps to maximize successful content retrieval across varied post formats. Extracted content was written to individual markdown files, organized into expert-named folders under research/linkedin-posts/. All files were manually verified for correct placement and content quality before being committed and pushed to GitHub.
Step 04: YouTube Transcript Collection : Designed an automated transcript pipeline using scripts/fetch_youtube_transcripts.py, built and refined with Cursor Agent. The script extracts video IDs from target URLs via regular expressions, generates a nested expert-folder structure under research/youtube-transcripts/, and saves each transcript as an individual markdown file following the naming convention Title [VideoID].md. All transcript files were verified for correct folder placement and content completeness before being committed and pushed to GitHub.
Step 05: Insights and Summaries : Produced structured summary files for both the LinkedIn posts and YouTube transcripts, covering expert name, content title, date published, duration (where available), URL, a 3 to 5 sentence summary of the main discussion, and the most important key insights. Stored in research/insights/ as linkedin-posts-summary.md and youtube-transcripts-summary.md.
Step 06: Supplementary Research Synthesis : Reviewed all collected material and produced two synthesized documents under research/other/: a consolidated content production framework drawing on the most actionable frameworks and workflows across all ten experts, and a structured playbook outline mapping the research themes into a logical, phase-by-phase content strategy sequence.
Step 07: README Update and Version Control : Updated the README to reflect the full project structure, collected content, tools, and process. Committed and pushed all changes to GitHub throughout development, maintaining a meaningful, incremental commit history rather than a single bulk push at the end.