RAG to Riches: Keyword Research for the AI Era

00:00 - RAG to Riches: Keyword Research for the AI Era
01:32 - Mark introduced, session promise: keyword research for AI and LLM era
03:31 - Mark starts slides, argues SEO fundamentals stay while execution changes
04:32 - PromptWatch Reddit chart shows how dependent ChatGPT is on Google results
06:17 - Why classic keyword lists fail for conversational, personalized LLM usage
07:23 - LLMs as universal intent decoder, replacing old “Google Fu” query skills
09:27 - Workflow overview: move from keywords to LLM style conversations and prompts
10:07 - Step 1: use personas in LLMs to turn seed keywords into real questions
12:05 - Step 2: use AlsoAsked to map follow up questions and decision journeys
15:04 - Step 3: predict which prompts are grounded and trigger real web searches
16:21 - Grounding explained: solved queries vs “query deserves freshness” behavior
18:50 - Using DevTools and Dan Petrovic’s classifiers to score grounding at scale
21:06 - Step 4: extract actual backend search queries LLMs send to search engines
23:04 - Build list of target queries, then do classic SEO gap and content work
25:20 - Multi site focus: get mentioned across many ranking pages, not just yours
29:20 - QueryFan.com teaser: free tool to automate personas, prompts and grounding
33:02 - Or: search volume is historical and weak, focus on intent and journeys
35:02 - Balancing LLM and classic SEO work, brand and entity optimization overlaps
40:30 - New KPIs: citations, mentions and sentiment vs pure traffic and rankings
45:00 - How to influence base models: digital PR, authority sites, community signals
48:10 - Handling negative SEO and LLM reputation issues, appeals and outranking
52:00 - Examples of brands with clear entities vs shallow “best X” listicle tactics
57:20 - Google vs ChatGPT usage: quick answers in Gemini vs deep chats in LLMs
60:01 - Final advice: use LLM intent to shape content, not giant FAQ dumps