Podcast Summary
The Best SEO Podcast: Defining the Future of Search with LLM Visibility™
Episode: SEO’s New Frontier With Duane Forrester
Host: Matthew Bertram
Guest: Duane Forrester
Date: February 23, 2026
Overview
In this masterclass episode, Matthew Bertram dives deep into the evolving world of SEO with industry veteran Duane Forrester. The discussion explores the seismic shifts in search, the integration of AI and large language models (LLMs), and the emergence of new skills and frameworks required for digital marketers. Bertram and Forrester break down advanced concepts like retrieval optimization, chunking, vector embeddings, and the limitations of old SEO metrics, offering rich insights for marketers racing to stay ahead in the AI-powered search era.
Key Discussion Points & Insights
1. The Evolution from Traditional SEO to the ‘Machine Layer’
- The internet’s second coming: AI is disrupting marketing much like the web’s original emergence ([00:17]).
- “It's 2026 and it's like the birth of the Internet has happened again.” – Bertram
- Duane Forrester’s background: From Caesar’s Palace and Microsoft Bing to Yext and schema.org, with deep involvement in AI ([01:23]).
- “I've been deep into AI since then and it's been eye opening, invigorating, frightening, exciting.”
Leveling Up: New Skillsets for SEOs
-
Modern SEO demands knowledge beyond keywords—embedding, chunking, semantic similarity, retrieval, and AI confidence scores ([04:00]-[05:25]).
- “Optimizing for retrieval, confidence and chunking structure... Not the same as optimizing landing pages for human conversion... Understanding vector embeddings and semantic similarity… not the same as running a keyword gap analysis.” – Bertram (quoting Forrester’s book)
-
The role of SEO must be elevated within organizations. Data-rich SEOs need full visibility and more decision-making power ([04:00]-[06:28]).
The Machine Layer Metaphor
- The book “The Machine Layer” guides SEOs from “high school to university”—leveling up is non-negotiable ([06:08]).
2. The Moving Target: Continuous Change and the Wild West
-
SEO in 2026 is as chaotic as its early days: "We are back in the Wild West." – Forrester ([14:42])
- The environment feels undefined and unstructured, requiring personal initiative and adaptability.
-
Industry hiring hasn’t caught up: Only 2% of SEO job postings require AI knowledge ([06:44]-[09:00]).
-
“You can't suck at SEO and excel with the AI environment. If your execution is not good, better or best... you are not going to perform in AI answers.” – Forrester ([08:40])
3. AI vs. Search Engines: How Discovery and Retrieval Work Now
LLMs are NOT Search Engines
-
LLMs don’t operate like databases or search engines ([37:20]-[39:16]):
- Search engines return a list, leaving humans to choose (agency). LLMs aim to give the definitive answer, shifting decision-making ([40:43]-[43:49]).
- “A search engine means I still have agency... the engine forcing the human to complete the final mile. LLMs...[say] I got you, bro. Here you go. Here's your answer.” – Forrester ([40:43])
-
Search is about information retrieval, but the execution is what creates the difference.
Indexing, Authority, and Citations
- LLMs may cite content not even indexed in traditional search (21% in some studies). There's little transparency on why or how results surface ([11:03]-[13:07]).
- “How do I know if I'm going to show up there (in AI answers)?”—Old tactics and KPIs are less relevant as each AI platform is opaque and evolving ([14:42]).
4. Structured Data, LLM.txt, and the Value of Providing Signals
- Structured data is still critical:
- “If your approach is, I don't need it... Thank you very much for making my life easier. I'm happy to have less competition.” – Forrester ([17:38])
- Debates on LLM.txt:
- LLM.txt lacks industry or platform backing; no major player has adopted it, and its value is unproven ([18:24]-[22:46]).
5. Modern KPIs: What Metrics Matter Now?
- Standard SEO KPIs (rankings, traffic) are quickly becoming obsolete ([23:51]-[24:33]):
- “The current KPIs of traffic, of keyword rankings are thrown out the window... They’re a point, but they're not enough.” – Bertram
- Proprietary data (semantic density, retrieval confidence, etc.) remains locked in LLM and search platforms ([24:33]).
- Marketers must manually track AI answer visibility using their own spreadsheets/monitoring ([57:27]-[58:36]).
6. Deep Dive: Semantic Density, Chunking, and Content Strategy
-
Semantic Density: LLMs evaluate the depth and completeness of knowledge in each content chunk ([25:57]-[27:02]).
- Content for LLMs must answer all relevant entity questions in focused, concise units, not spread thinly for human reading flow ([26:13]-[27:02]).
- “Do you provide everything I need within that chunk to understand that topic, question, that entity?” – Forrester ([26:13])
-
Human hooks and emotional appeals are largely ignored by AI—objectivity, factual accuracy, and broad consensus matter most ([29:00]-[31:47]).
- Example: Amazon’s AI summaries reflect contradictory crowd opinions, not nuanced judgments ([31:47]).
-
Every platform and model values different things; writers must adapt for each (news vs. niche topics, etc.) ([34:09]-[36:33]).
7. Vector Embeddings: The New Core Concept
- Explanation ([49:46]):
- “The embedding is a mathematical representation… that number exists in a three-dimensional space. The vector is directional… how close is that number to the number created for the query.”
- “You are attempting to chase a meteor hurtling through space and you want to come up as close behind it, to land on it as possible.”
- SEOs need to align content to sit as close as possible in vector space to likely queries—this is a dynamic, ever-shifting target ([49:46]-[54:24]).
- For deeper understanding, Forrester recommends researching Weaviate, Pinecone, Supabase, and more.
8. The Divide Ahead: Math-Driven SEO vs. Checklist SEO
-
Much of the current field is content-focused or checklist-driven; the LLM era demands mathematical, data-driven understanding ([54:24]-[56:04]).
- “Your future is all about math and if you don’t like it... you're about to take a left hand turn off the highway...” – Forrester
-
No prescriptive checklists exist for LLM optimization due to lack of transparency from platforms ([56:04]-[57:25]).
- Manual tracking and analysis using AI tools on your collected data is recommended.
9. Adapting and Learning: The Pace and Community Need
- AI/LLM platform updates come twice as fast as Google algorithm updates ([62:09]).
- “Don’t worry if you can’t keep up with every single thing—what’s important is that you know how to get back into that pipeline...” – Forrester ([62:09])
- Community sharing and experimentation will define the new frameworks and best practices ([59:36]-[62:09]).
Notable Quotes & Moments
- “[AI in marketing is] eye-opening, invigorating, frightening, exciting.” – Forrester ([01:23])
- “...there's an additional layer on top of, of, of good SEO. So it AI doesn't protect you from bad SEO that you have tech debt you got to make up.” – Bertram ([04:00])
- “You can't suck at SEO and excel with the AI environment.” – Forrester ([08:40])
- “We are back in the Wild West. That's where SEO started.” – Forrester ([14:42])
- “If your execution of SEO is not... best, if it's mid level or lower, you are not going to perform in AI answers.” – Forrester ([08:40])
- “Your future is all about math and if you don't like it... a whole lot of the industry is about to take a left-hand turn.” – Forrester ([54:46])
- “No prescriptive checklists exist for LLM optimization.” – Forrester ([56:04])
- “...you are attempting to chase a meteor hurtling through space and you want to come up as close as possible.” – Forrester on vector embeddings ([49:46])
- “I’m sick of using all these different tools and then trying to triangulate in my head what the answer is.” – Bertram ([59:36])
- “Updates at platforms like ChatGPT and Claude... are more than twice as fast as what we’re used to as SEOs.” – Forrester ([62:09])
Key Timestamps
- 00:17 – Setting the stage: AI revolutionizes marketing and search.
- 01:23 – Duane’s background and schema.org, Yext, AI involvement.
- 04:00 – Skills gap in modern SEO; AI requires a new, advanced toolkit.
- 06:28 – Meta-level: Leveling up from traditional to machine-layer SEO.
- 14:42 – The Wild West analogy; uncertainty in SEO's new frontier.
- 17:38 – On structured data's ongoing importance.
- 18:24 – The value (or lack thereof) of LLM.txt.
- 23:51 – The obsolescence of traditional SEO KPIs.
- 25:57 – Semantic density vs. semantic search.
- 29:00-31:47 – Why emotional content is ignored by LLMs.
- 34:09 – How resource allocation differs across verticals.
- 37:20-40:43 – Difference between search engines and LLMs.
- 49:46 – Vector embeddings, explained.
- 54:24 – Division in the SEO industry: math-driven vs. checklist approaches.
- 57:25 – Manual tracking in the AI/LLM era.
- 62:09 – The pace of change and need for constant learning.
Final Thoughts & Guest Plugs
- Duane Forrester’s frameworks (some free) and book “The Machine Layer” are available at duaneforrester.com.
- The need for community, open sharing, and rethinking professional development is vital—practitioners must reinvent how they learn, execute, and measure SEO in the LLM era.
- Keep experimenting, share findings, and avoid relying on outdated tools and metrics.
This thorough, future-forward episode offers both practical advice and philosophical context for anyone looking to lead in the AI-driven search landscape.
