From SEO to AI Search Optimisation - CourtneyR.dev's 2nd Brain

From SEO to AI Search Optimisation ![rw-book-cover](https://img.youtube.com/vi/trOKoG28A2s/maxresdefault.jpg) 🔗 **Source:** [From SEO to AI Search Optimisation](https://www.youtube.com/watch?v=trOKoG28A2s) > [!note]+ Document Notes > tags: #seo #ai-search #search-optimization #wordpress #schema-markup #content-strategy #search-engines #llm > author: Alex Moss > source: WP26 Event > reference: https://www.youtube.com/watch?v=trOKoG28A2s > date created: 2024-12-19 10:30 > > # 🏷️ Title: From SEO to AI Search Optimisation >  ## 🧰 Action Inbox - [H1] Audit existing content for machine readability and eliminate fluff — editorial standards critical for AI discovery #content-strategy - [H2] Enable Yoast schema aggregation endpoint for site-wide structured data mapping — single API call for LLM ingestion #schema-markup - [H3] Test AI tools on your own site to understand how machines interpret your content — hands-on evaluation essential #ai-search ## 🧭 Table of Contents - [[#📝 Summary]] - [[#📖 Main Content]] - [[#🗒️ Notes]] - [[#📚 References]] - [[#🗣️ Paraphrase]] - [[#📰 Newsletter]] - [[#✅ Action Items]] - [[#🧪 Hypotheses]] - [[#📊 Methodology]] - [[#❓ Questions]] - [[#🔮 Future Research]] - [[#💡 Implications]] - [[#🧠 Learning Styles Assessment]] - [[#📊 Accessibility Assessment]] - [[#🌍 DEIB Assessment]] - [[#✅ Logical Validity]] - [[#🛡️ DARVO Analysis]] - [[#🎯 Conclusion]] ### 🔗 References - [Yoast Schema Aggregation](https://yoast.com) - [Microsoft NL Web Team](https://microsoft.com) - [Google Search Central](https://developers.google.com/search) ## 📝 Summary - AI search engines are transforming how content gets discovered, moving from keyword lists to conversational, contextual answers - Good traditional SEO practices remain foundational, but content must now be optimized for both human and machine readability - New technologies like schema aggregation endpoints enable LLMs to understand entire site structures through single API calls - Success metrics are shifting from clicks/impressions to discoverability and brand sentiment across AI platforms - The core opportunity lies in bridging marketing, product, and development teams through enhanced SEO practices ## 📖 Main Content The search landscape has fundamentally shifted from the simple "10 blue links" of early Google to AI-powered conversational interfaces that synthesize answers from multiple sources. Alex Moss, Principal SEO at Yoast, explains how this evolution requires updating our optimization strategies while maintaining core SEO principles. The transformation involves moving from keyword-focused optimization to what's being called Generative Engine Optimization (GEO) - though multiple acronyms exist (AEO, AIO, LLMO). The key insight is that LLMs now act as additional audience segments that need to discover, understand, and cite your content to provide accurate answers to users. Modern AI search relies on "grounding" - the validation of information from your website against other web sources to minimize hallucinations. This requires content that's not just human-readable but optimized for machine parsing through proper structure, schema markup, and clear formatting. Commercial applications are advancing rapidly, with OpenAI's Agentic Commerce Protocol enabling purchases directly within chat interfaces, and Google's Universal Commerce Protocol handling the entire buyer journey from discovery to logistics without leaving the AI environment. ## 🗒️ Notes #### ⏱️ [00:00](https://youtu.be/trOKoG28A2s?t=0) — [H1] Introduction to AI Search Transformation - Key idea: Search has evolved from simple link lists to conversational, contextual AI-powered answers - Evidence: "LLMs disrupted the native search experience... it's now about retrieval and generation" #### ⏱️ [03:45](https://youtu.be/trOKoG28A2s?t=225) — [H2] The Evolution from Links to Agents - Key idea: Personal AI agents do extensive background work beyond human input to provide personalized results - Evidence: "The green part is the human input and everything else is the agents input... they're doing way more work" #### ⏱️ [06:30](https://youtu.be/trOKoG28A2s?t=390) — [H1] Generative Engine Optimization Defined - Key idea: GEO optimizes content for AI search environments to make it discoverable and authoritative - Evidence: "Optimizes content for generative AI search environments... to make it discoverable, trustworthy, and authoritative" #### ⏱️ [09:15](https://youtu.be/trOKoG28A2s?t=555) — [H2] Microsoft's GEO Framework - Key idea: Microsoft published comprehensive research on GEO methodology and implementation - Evidence: "Microsoft... presented a white paper in January... added AEO as well as GEO" #### ⏱️ [11:45](https://youtu.be/trOKoG28A2s?t=705) — [H1] Grounding and Content Validation - Key idea: AI systems validate website information against multiple sources to ensure accuracy - Evidence: "Grounding... is basically the validation of information from your website corroborated with other sources" #### ⏱️ [14:20](https://youtu.be/trOKoG28A2s?t=860) — [H2] Multimodal Search Capabilities - Key idea: AI can now process combinations of text, images, video, and documents simultaneously - Evidence: "Gemini embedding 2... combination of using text, images, video and documents... can combine them all in the same model" #### ⏱️ [17:10](https://youtu.be/trOKoG28A2s?t=1030) — [H1] Agentic Commerce Integration - Key idea: AI agents can complete entire purchase journeys without users leaving the chat interface - Evidence: "Agentic commerce protocol... lets you inside chat GPT do a search you find a product and then you can purchase it" #### ⏱️ [19:45](https://youtu.be/trOKoG28A2s?t=1185) — [H2] Web Model Control Protocol - Key idea: Google's MCP enables AI agents to interact directly with website forms and elements - Evidence: "Web MCP... can identify where it is, what the inputs are for... populate it and submit that form on your behalf" #### ⏱️ [22:30](https://youtu.be/trOKoG28A2s?t=1350) — [H1] Good SEO Equals Good GEO - Key idea: Traditional SEO best practices correlate strongly with AI search optimization success - Evidence: "Good SEO is good GEO because they correlate... sites doing well in LLM... not doing well in native SERPs" #### ⏱️ [25:15](https://youtu.be/trOKoG28A2s?t=1515) — [H2] Editorial Standards Enhancement - Key idea: Content quality standards must improve as mediocre content can no longer survive AI filtering - Evidence: "No longer can mediocre content survive... making content for content's sake isn't what's going to win" #### ⏱️ [28:40](https://youtu.be/trOKoG28A2s?t=1720) — [H1] Data No-Snippet Controls - Key idea: New HTML attributes allow precise control over what content appears in AI summaries - Evidence: "Data no snippet tag... allows you to have precision content controls... stay in the index and be ranked but excludes it from snippets" #### ⏱️ [32:10](https://youtu.be/trOKoG28A2s?t=1930) — [H2] Brand Sentiment and Success Metrics - Key idea: SEO metrics are expanding beyond clicks to include brand visibility across AI platforms - Evidence: "Brand visibility and success metrics... goes beyond just the old metrics that we used to have" #### ⏱️ [35:25](https://youtu.be/trOKoG28A2s?t=2125) — [H1] Markdown vs HTML Debate - Key idea: AI agents prefer markdown format but search engines want single HTML versions - Evidence: "Agents prefer markdown to markup... the search engines do not want you to have an alternate version" #### ⏱️ [38:50](https://youtu.be/trOKoG28A2s?t=2330) — [H2] Schema Aggregation Innovation - Key idea: Yoast's schema map provides entire site structure data through single API endpoint - Evidence: "Schema aggregation... provides standardized... map of all of your entities via one single API call" #### ⏱️ [42:15](https://youtu.be/trOKoG28A2s?t=2535) — [H1] Future SEO Opportunities - Key idea: SEOs will become more valuable as bridges between marketing, product, and development teams - Evidence: "SEOs may become a lot more valuable... be the key and the bridge between all of those things" ## 📚 References - [Yoast SEO Plugin](https://yoast.com/wordpress/plugins/seo/) - [Schema.org Documentation](https://schema.org/) - [Google Search Central Live](https://developers.google.com/search/events) - [Microsoft NL Web Framework](https://nlweb.org/) - [Cloudflare Crawl Endpoint](https://developers.cloudflare.com/) ## 🗣️ Paraphrase - Search has transformed from showing lists of website links to AI assistants providing conversational answers by synthesizing information from multiple sources - Traditional SEO fundamentals remain important, but content must now be structured for both human readers and AI systems that need to understand and cite it accurately - New tools like schema aggregation endpoints allow AI to understand entire websites through single data calls rather than crawling individual pages ## 📰 Newsletter **AI Search is Here: What WordPress Sites Need to Know** The search landscape just shifted dramatically. Instead of users clicking through lists of links, AI assistants now provide direct answers by reading and synthesizing content from across the web. For WordPress site owners, this means both opportunity and urgency. **The Good News**: Your existing SEO knowledge still matters. Sites that rank well in Google tend to perform well in AI search too. The fundamentals—quality content, proper structure, and clear information—remain crucial. **The New Reality**: Your content needs to work for two audiences now—humans AND AI systems. This means using proper headings, writing clear summaries, citing sources, and ensuring your content is "machine readable" through structured data. **Quick Wins for WordPress**: - Update to latest Yoast SEO and enable schema aggregation - Audit your content—eliminate fluff and ensure every page clearly states its purpose - Use proper headings, lists, and formatting (the basics many still skip) - Add schema markup for your content types The shift from "SEO" to "AI Search Optimization" isn't about throwing out everything you know—it's about doing the fundamentals better and thinking about discoverability, not just rankings. Resources: [Yoast SEO](https://yoast.com/wordpress/plugins/seo/), [Schema Markup Guide](https://developers.google.com/search/docs/appearance/structured-data), [WordPress SEO Learning Path](https://learn.wordpress.org/course/wordpress-seo/) ## ✅ Action Items #### WordPress - Core: Integrate AI search optimization guidelines into WordPress SEO documentation and best practices - Plugins: Evaluate plugin ecosystem readiness for schema aggregation and structured data enhancement - Themes: Develop theme standards that support machine-readable content structure by default - Documentation: Create guides bridging traditional SEO and AI search optimization for WordPress users - Community: Establish working group on AI search optimization best practices for WordPress ecosystem #### OSPO - Policy & Governance: Develop guidelines for open source projects optimizing for AI discovery and citation - Community Engagement: Support schema.org and other web standards critical for AI search functionality - Developer Support: Provide training on structured data implementation and AI-readable content creation #### Community - Maintainers: Document how AI systems discover and cite open source project information - Contributors: Train contributors on creating machine-readable documentation and structured content - Sponsors: Measure brand visibility and sentiment across AI platforms alongside traditional metrics #### DevRel - Developer advocacy: Present on AI search optimization at WordCamps and developer conferences - Education: Create workshops on schema markup implementation and structured data best practices - Developer experience: Improve documentation structure for better AI system comprehension - Content strategy: Audit developer resources for machine readability and structured data compliance ## 🧪 Hypotheses - Websites with comprehensive schema markup will achieve higher citation rates in AI search results than those relying solely on HTML structure - Traditional SEO ranking factors correlate positively with AI search visibility across different language model platforms - Sites optimized for machine readability will see improved user engagement as AI systems provide more accurate summaries and recommendations - The markdown versus HTML debate will be resolved by AI systems developing better HTML parsing capabilities rather than requiring alternate formats ### 📊 Methodology - Comparative analysis of AI citation rates between schema-enhanced and basic HTML websites across multiple LLM platforms - Correlation study between traditional search rankings and AI search visibility using tools like Yoast AI Business Insights - A/B testing of content formats (markdown vs. HTML) on AI system comprehension and accuracy metrics - Longitudinal study of brand sentiment changes across AI platforms following structured data implementation ### ❓ Questions - How will the proliferation of AI search affect organic traffic patterns and business models dependent on website visits? - What happens to long-tail keyword strategies when AI systems synthesize answers from multiple sources rather than directing users to specific pages? - How can smaller websites compete for AI visibility against large-scale content producers with extensive structured data resources? ### 🔮 Future Research - Development of standardized AI search optimization auditing tools and metrics beyond current brand sentiment tracking - Integration patterns between schema aggregation endpoints and emerging AI agent frameworks for commerce and content discovery - Evolution of search engine guidelines as traditional SERPs and AI-generated answers compete for user attention and advertiser revenue ### 💡 Implications - Practice: SEOs must expand skillsets to include structured data architecture and AI system optimization alongside traditional ranking factors - Policy: Web standards organizations may need governance frameworks for how AI systems cite and attribute source content - Education: WordPress training programs require updates to include AI search optimization as core competency rather than advanced topic - Funding: Investment in structured data infrastructure and content quality becomes essential for digital marketing ROI ### 🧠 Learning Styles Assessment Reading level estimate: 12th grade (technical concepts explained accessibly) #### 👀 Visual learners - Present: Schema visualization tool demonstration, search evolution timeline graphics - Improve: Flowcharts showing AI search decision trees, before/after content optimization examples #### 🔊 Aural learners - Present: Full video presentation with clear narration and Q&A session - Improve: Podcast-style deep dive on specific tools like schema aggregation #### 🗣️ Verbal learners - Present: Clear definitions of technical terms (GEO, grounding, schema aggregation), real-world analogies - Improve: Case study narratives showing transformation journeys from traditional to AI-optimized SEO #### 🧩 Kinesthetic learners - Present: Step-by-step Yoast schema aggregation setup instructions - Improve: Interactive workshop format with hands-on schema markup creation and testing #### 🧠 Logical learners - Present: Correlation data between traditional SEO and AI search performance, structured methodology comparisons - Improve: Decision trees for choosing optimization strategies, ROI calculation frameworks #### 🧑‍🤝‍🧑 Social learners - Present: Q&A interaction, community discussion prompts, expert networking opportunities - Improve: Peer review sessions for content optimization, collaborative schema markup projects #### 🧘‍♀️ Solitary learners - Present: Comprehensive resource lists, self-directed learning paths through linked materials - Improve: Self-assessment checklists for AI search readiness, personal optimization workbooks #### 🌿 Naturalistic learners - Present: Web ecosystem evolution analogies, organic search environment changes - Improve: Sustainability implications of AI search on web publishing ecosystems ### 📊 Accessibility Assessment #### Screen reader semantics - Present: Well-structured presentation with clear headings and logical flow - Improve: Detailed alt text for schema visualization demos, ARIA labels for interactive elements #### Keyboard navigation - Present: Standard video controls accessible via keyboard - Improve: Timestamped navigation menu, skip-to-content links for long-form presentation #### Contrast and readability - Present: High contrast slides with readable fonts - Improve: Captioned technical diagrams, color-blind friendly schema visualizations #### Text resizing and reflow - Present: --- ## Full Transcript [00:06] Welcome back everybody. We're live again [00:09] for the second session of the WP26 [00:13] uh event and we're going to be talking [00:15] about SEO and AI search optimization. in [00:19] this session with uh Alex Moss who is a [00:24] principal SEO expert at the coast and [00:26] has like more than 15 years of [00:29] experience across agency in in-house and [00:31] freelancer roles uh specialized in [00:34] technical uh SEO and has worked with [00:38] WordPress since uh 2010 has built a lot [00:42] of plugins a lot of themes and large [00:45] scale SEO strategies for worldwide [00:49] organizations. [00:50] Uh Alex will give us an overview of how [00:53] how AI is transforming the way we work [00:56] with content and he's going to share [00:59] ways we can respond and explore how it [01:01] might shape the future of the web uh and [01:04] content discovery uh in the next years [01:08] to come. So, welcome Alec. Let me pull [01:11] you uh up. There you go. Hello. [01:15] >> Hi. How you doing? [01:16] >> I'm doing really well. really excited to [01:19] have you here. We've we like ran a [01:21] little bit late with no so uh without [01:23] further ado, over to you, Alex. [01:26] >> Thank you. Thank you very much and hello [01:29] everyone. Thank you. Um thank you for [01:31] coming. So let me do presenter of you. [01:33] Um just as a word of warning, I'm going [01:35] to be doing this offline in presenter of [01:37] view because Microsoft 365 and Crowdcast [01:40] don't really get along. So we'll have to [01:42] make do. So I'm very sorry. I see Tacos. [01:44] Woo, let's go. So, let me listen to Taco [01:47] because that's always the best way to do [01:49] So, I'm going to talk to you about how [01:52] we've gone from SEO to AI search [01:55] optimization. Um, because a lot's [01:57] changed recently and it may be worth um [01:59] just refreshing everyone's mind about [02:01] what's happening. Not all the AI stuff [02:04] that's happening. I know that null was [02:05] talking a lot about how you can scale [02:07] and improve your workflows, but I want [02:09] to talk about things from a search [02:10] perspective. Um, as a little bit of [02:12] background, I know Petia did a wonderful [02:14] introduction of me. Um, I'm principal [02:16] SEO at Yoast. I'm also a co-founder and [02:18] director of a small agency called Fire [02:20] Cask, which is based in Manchester in [02:22] the UK. Um, we do organic marketing, [02:25] omni channel stuff, and as well as that, [02:27] of course, um, development and design in [02:29] our favorite CMS. Um, but I've actually [02:32] been doing SEO since 1999, which kind of [02:35] shows my age, also shows the passion [02:37] that I've got in optimizing websites in [02:39] order for them to be discoverable, even [02:41] though I didn't know that's what I was [02:42] doing at the time. And back in 1999 to [02:45] 2002, life was so much simpler because [02:48] we just had this list of 10 links. You [02:51] did your own research. You saw and [02:52] decided what was best for you, and then [02:55] you decided to go away with that [02:56] information or buy that product or take [02:58] on that service. Now, as the years went [03:01] on, um you saw that Google started [03:03] buying up more assets like introduced [03:05] Adwords, they bought YouTube, and they [03:08] brought it all into the search results [03:09] page. And over the years, you can see [03:11] here that the green organic part has [03:13] been getting less and less so and been [03:16] taking up more real estate, which is a [03:18] bit uh made it a bit annoying for us [03:20] SEOs, but made us a bit more valuable [03:22] over time. But then a lot has kind of [03:25] happened specifically in the last two [03:27] years as I'm sure you might have [03:28] noticed. Um, and one thing from my point [03:31] of view is that LLM's disrupted the [03:34] native search experience. In a way, it's [03:36] kind of gone full circle. It's kind of [03:38] looking a lot more simpler than it used [03:40] to be. And it wasn't just, you know, [03:41] lots of flashy assets all around that [03:44] may confuse you as you're doing your own [03:46] research. Instead, it's no longer a list [03:48] of links. It's now about retrieval and [03:50] generation. And it kind of combines that [03:52] experience to give you a really [03:54] conversational and contextual [03:56] synthesized answer. But more [03:58] importantly, it's personalized and over [04:00] time has come with memory and more [04:02] context about you as a person. So for [04:04] example, I know that Gemini connects [04:06] with Gmail. I've had Gmail for I don't [04:09] know how long it is. It was in beta for [04:11] about three years, but I was definitely [04:13] the first one million who got an email [04:15] there. Um, so it's got about 20 years [04:17] worth of emails in there. So it knows a [04:19] lot about me. Um, but what's also more [04:22] important is that this flow has changed. [04:24] It's gone from, you know, a keyword, you [04:27] get the answers, that's not perfect. [04:29] Rinse and repeat until you get the [04:31] research done from what you want. Now, [04:33] it's all about your user intent, taking [04:36] that result and adding additional [04:37] context to hone down exactly what you [04:40] want as an end result. Now, behind the [04:42] scenes, this is kind of what a personal [04:44] LLM user agent looks like. It takes the [04:47] user inputs. It com it it connects to an [04:49] LLM. It then access it then assesses and [04:51] engages personal context network of [04:54] agenda and all the stuff that you kind [04:55] of know already. But what's more [04:57] important here is that the green part is [04:59] the human input and then everything else [05:02] is the agents input into what's going [05:05] on. So underneath the hood, they're [05:07] doing way more work than what we were [05:10] doing in the first place. But what's [05:11] more important from a search perspective [05:14] is that we're going from clicks to [05:16] decisions and selections through agents, [05:20] right? And that's kind of been given a [05:22] lot of debate in search about what term [05:26] that may be. And here we'll enter geo [05:30] and that is short for generative engine [05:32] optimization. And here's a definition. [05:35] So I'll read it out. optimizes content [05:37] for generative AI search environments [05:39] like a LLM powered engines to make it [05:42] things discoverable, trustworthy, [05:45] unauthoritative. Right? So search is no [05:47] longer this list of blue links. It [05:50] recaps, it summarizes answers and gives [05:52] you personalized results delivered by [05:54] assistance. Now GIO is an acronym that [05:57] has been adopted um by quite a few SEOs, [06:01] but it's up for debate because there's [06:02] so many out there. But one company [06:04] that's been really honing in on the term [06:06] go and AO is Microsoft. They actually um [06:10] presented um a white paper in January um [06:14] and it goes back to October as well when [06:17] they were started writing about GA uh [06:19] GIO as a specifics. But here it gives [06:21] you lots of information and this is one [06:23] screenshot um of how at least from [06:26] Bing's point of view you and an LLM [06:29] works behind the scenes in order to give [06:31] you the best possible answers. But one [06:34] thing I do notice here is that they've [06:35] added AEO as well as GEO and AEO is [06:40] answer engine optimization. [06:42] And again it's just another acronym. In [06:44] fact, there's loads of acronyms about [06:47] there's GIO, AEO, AIO, AI, SEO, LLM O, [06:51] LLM, SEO, and VO. VO is vector embedding [06:55] optimization. Um there's just so many. [06:58] Um but one thing that's common between [07:00] all of them is is a nice little diagram [07:03] that John Müller presented in um Google [07:06] search central live in Zoric which is [07:08] grounding which is basically the [07:10] validation of information from your [07:13] website corroborated with other sources [07:15] on the web. It uses that validation to [07:18] ensure that the answer is the most [07:19] accurate and has the least amount of [07:21] hallucinations. [07:23] um it shapes content so LLMs can pass it [07:26] with minimal guesswork and that also [07:29] means that we have to get on board as an [07:31] SEO to make sure that things like schema [07:33] headings modular content in chunks are [07:36] all optimized to be not just human [07:38] readable but machine readable and that [07:41] goes a step further now we have [07:43] multimodal search and this is a good [07:44] example that Google gave where you can [07:47] select a picture you can take a picture [07:49] of a few books and say if I enjoyed [07:51] these what are some similar books and [07:53] you can see it working. Now, [07:56] when it does that, it takes those books, [07:58] it understands everything about those [08:00] books, it then presents recommendations [08:02] based on the combination of all those [08:03] books, and then it gives you information [08:05] about them presented in a nice little [08:07] list that you can then do more research [08:09] on. But what does that mean for search? [08:13] It means that technically [08:15] the agent is an additional p. It's an [08:18] additional audience that audience [08:20] segment that we kind of have to attract [08:22] even not more so than the human [08:24] nowadays. So much so that Gemini only [08:27] two days ago released um Gemini [08:29] embedding 2 which is a combination of [08:33] using text, images, video and documents [08:36] such as PDFs can combine them all in the [08:39] same model and understand all of those [08:41] multimodal entities and elements and the [08:43] relationships between them. And that [08:45] means it's actually really important. [08:47] Things go way beyond text now. And now [08:49] your short video on YouTube can now be [08:52] understood in the same way that reading [08:54] a book can. And that's also really [08:56] important because all of these things [08:58] can combine and these entities can have [09:00] relationships with each other. Um [09:02] developers can search across all of the [09:04] platforms using one single system um [09:06] instead of managing all these separate [09:08] tools. And I think this is um a trend [09:10] that's going to continue about this [09:12] consolidation of tools into one unifying [09:15] very convenient thing. Now as well as [09:18] that we've got commerce and in commerce [09:21] specifically we've had in September [09:23] Agentic commerce protocol which was [09:26] introduced by open AI um which kind of [09:29] lets you inside chat GPT do a search you [09:32] find a product and then you can purchase [09:33] it without leaving that chat session and [09:37] that's really important because that [09:38] means that no longer does a person have [09:40] to go to a website um make a decision go [09:44] through a checkout add all your [09:46] information and then go through. It's [09:47] seamless from the customer's point of [09:49] view. That's really really convenient. [09:51] Now, UCP was um introduced in January um [09:56] by Google and that's an open framework [09:58] that's still yet to be well started [10:00] rolling out in a few places like Walmart [10:03] um and Wayfair and Etsy using it as well [10:06] where it doesn't just take the [10:07] transactional element. It takes the [10:09] whole journey from discovery all the way [10:11] to logistics. Um, and that's also really [10:14] important because you soon you won't [10:16] need to leave AI mode to understand what [10:18] you want, when you're going to get it, [10:20] where it's going to be, and customer [10:22] service as well. It also will connect to [10:24] that. If anything, the agents completing [10:26] the entire buyer journey, and that's [10:28] very important for an SEO, and we'll go [10:31] on to that in a sec. Google have also [10:34] released the web MCP, and it's currently [10:36] there um for early preview. They've got [10:39] a Discord um community as well. There's [10:41] a mailing list where you can go in and [10:43] see what's this about. But really, [10:45] that's kind of like the new SEO for [10:46] agents specifically. They go under the [10:49] hood. They look at HTML and they look at [10:51] elements within HTML to understand the [10:54] intent of those places. So, as an [10:56] example, you might use a contact form. [10:58] It can identify where it is, what the [11:00] inputs are for, and as well as that, it [11:02] can populate it and submit that form on [11:05] your behalf. Those are all done using [11:07] imperative JavaScript tools, API, and [11:10] the declarative HTML form attributes all [11:14] as tools. And this is interesting [11:16] because it's actually model agnostic as [11:19] well. So any LLM can use this right now. [11:23] But what do we learn from all of this? [11:24] That in fact with GIO and LLMO or [11:28] whatever the acronym is, good SEO is [11:31] good GEO. Now, this is a slide or a [11:34] pause of a slide from um Danny Sullivan [11:36] who gave the keynote in Word Camp US [11:38] last year. And the same still applies [11:41] today and good SEO is good GEO because [11:44] they correlate. And I'd like to [11:45] challenge by saying that I'm not finding [11:48] um many or if any sites that are doing [11:51] really well, say in an LLM such as Chat [11:54] GPT or Perplexity or Claude that's not [11:56] doing well in the native SERs that are [11:59] still there and vice versa. So really [12:02] what should we be thinking of as an SEO? [12:04] What should SEOs be focusing on this [12:06] year? And I've split this into four key [12:09] points. The first is editorial [12:11] standards. Next is brand sentiment. Then [12:14] success metrics which is going to be [12:15] confusing. And lastly data integrity. So [12:18] if we go through editorial standards, [12:21] it's something that needs to be [12:22] improved. And I actually think this is [12:23] for the good of the web a really really [12:26] positive thing. No longer can mediocre [12:29] content survive. Um, and making content [12:32] for content's sake isn't what's going to [12:34] win. And it hasn't in the past, but it's [12:36] going to happen even faster today. You [12:39] just want to make unique human-driven [12:41] content that wins, but it also has to be [12:43] machine readable. So doing things like [12:46] introducing a summary at the beginning, [12:48] having a concise conclusion at the end, [12:50] making the most important things in the [12:52] top and bottom 20% um of your content, [12:55] making things like good formatting, [12:57] right? It says use good formatting. It's [12:59] very interesting because that's always [13:00] been there. And I know that you [13:01] WordPress afficionados out there are [13:03] going to be like, "This is obvious." But [13:05] weirdly, from an editorial point of [13:06] view, people don't do the basics such as [13:09] entering headings or increasing tables [13:12] um or making lists and making sure that [13:15] evidence and citations are very clearly [13:18] there and done so in a very very nice [13:21] formatted and structured manner so that [13:23] a machine as well as a human can [13:25] interpret that. Well, um, you've got to [13:27] site your sources. Provide evidence of [13:30] whatever you're writing about. Um, and [13:32] you don't just focus just on keywords [13:34] anymore. This isn't about keyword [13:35] stuffing or minimum words, anything like [13:38] that. What you need to focus on is is [13:40] this going to be helpful? Is it [13:42] additive? Does a user find it useful? [13:45] And is is a machine going to find it [13:47] useful enough to site it and bring it [13:49] back to the human. Um, and as well as [13:52] that, I mean, there's the obvious of [13:54] using schema. And I know a really really [13:56] popular plug-in for WordPress that does [13:58] all of that. So check that out. Now [14:02] there's also interestingly um today that [14:04] was also shared by Krishna Madavan from [14:07] Microsoft. He's currently at the Friends [14:08] of Search conference today in Amsterdam. [14:11] um and he shared more information about [14:13] the data no snippet tag which allows you [14:16] to have precision content controls and [14:19] it's HTML attribute where you can stay [14:22] in the index and be ranked but it [14:25] excludes it from snippets such as AI [14:28] summaries AI overviews and AI mode. Now [14:31] it may be questionable and a business [14:33] case for you as to whether you want to [14:35] do that or not. But there are a few use [14:38] cases that he shared such as protecting [14:41] premium or payable content, managing [14:43] brand reputation, removing that legal [14:45] mumbo jumbo like terms and conditions [14:47] that serve no real purpose to even be [14:50] looked at and brought back to the user [14:52] and excluding sponsored or promotional [14:54] content. There are use cases, but do [14:56] handle these things with care. The [14:59] second thing is discoverability is the [15:01] additional goal. It's not just about [15:03] rankings or keywords. It's about [15:05] discoverability and assistance to an [15:07] agent to try and synthesize that answer [15:10] back to you. And whilst our goalposts [15:11] have changed, we therefore have to shift [15:13] with that change as well. It's a [15:15] structural shift and how people find [15:17] that trusted information. Now again, the [15:21] kind of the old hat and the new hat, [15:22] they're very similar. You've got content [15:25] is now requires more context. keywords [15:28] isn't just about that. It's about intent [15:30] and so on and so forth. But something [15:31] that's really interesting is EAT, which [15:34] um if you should read into it a bit [15:36] more, it's about experience um [15:39] experience. Oh, I forgot the other E. [15:42] Experience doesn't matter. Authority and [15:44] trust. And in that um it kind of [15:47] provides a a framework for how ethics [15:50] should be in terms of authority um on [15:53] your site. make sure that everything is [15:56] backed up by a professional i.e. you or [15:58] the brand and that brings everything [16:00] back to make sure it's validated by [16:02] whatever's looking into it. [16:05] A good version of brand visibility and [16:07] success metrics and sentiment is [16:10] something of tools that are coming out. [16:12] So you've got the big ones like profound [16:14] and such. Yoast have a very very [16:17] affordable way of showing how a brand's [16:20] sentiment is shown and visibility is [16:22] shown in different LLMs. And at the [16:24] moment we've got um Open AAI, [16:27] Perplexity, Claude and Google um all in [16:30] the same hood that looks at everything [16:32] and brings it all together and sees how [16:34] your brand is actually seen online. It's [16:37] something quite important and it goes [16:38] beyond just the old metrics that we used [16:41] to have. But even there Google are [16:43] announcing new reporting to search [16:45] console. They've introduced in the last [16:47] three months or announced keyword [16:49] grouping, brand and non brand query [16:51] filters and AI and AI mode metrics. But [16:55] at the moment they've only really and in [16:57] the last 24 hours um brought out the [17:00] brand and non-brand query filter which [17:01] you can now go inside in search console [17:04] and have a look. But really it's just [17:06] doing something that you could have [17:07] filtered out in the first place. Bing [17:10] have kind of tried to do a little bit [17:11] similar. Um they've brought something [17:14] like grounding queries in an AI [17:16] performance report. But again a the [17:18] grounding queries are important because [17:20] they're keywords that the agent is using [17:23] in order to search. So it's still [17:25] important that whilst the keyword is [17:27] something you need to focus on, it has [17:29] to be in the right context and it shows [17:32] what the grounding queries are and the [17:34] pages that they relate to. And when [17:37] things are machine readable, it's good [17:39] to think about what format that might [17:41] be. And one debate that's been around [17:44] quite a bit is the markdown debate. Um, [17:46] agents, as we all know, prefer markdown [17:48] to mark up. They prefer the concise [17:51] structure. They don't need all the [17:52] visuals whilst they can read everything. [17:54] It's being supported by Cloudflare. Um, [17:57] Matt Mullenweg wrote a post himself [17:59] about it. That's also gone into not just [18:01] WordPress.com, but WordPress.org now [18:03] have um a real alternate markdown. And [18:06] they've also introduced a plugin as well [18:08] as Yostastalk as well who's also brought [18:10] out a plugin that shows an alternate [18:12] version of a page in markdown and you [18:14] should definitely check that out um [18:16] because I think it's going to be um a [18:18] bit more popular as time goes on. [18:20] Interestingly, the two people who are [18:22] not supporting this are John Müller from [18:24] Google and Fris Canal from Bing. The [18:27] search engines do not want you to have [18:28] an alternate version. My guess is that [18:31] they don't want to read two versions and [18:33] then have to connect them together and [18:34] they're just telling us to have one. But [18:36] I guess they may not always be exactly [18:38] defining what goes forward in the [18:40] future. Um, and that brings on to how [18:44] not just how things are structured, but [18:47] data and entities and structured data. [18:50] And I personally believe I've always [18:52] been a fan of schema um and structured [18:54] data in the past, but as time's gone on, [18:56] I'm believing more so that it's even [18:57] more important. ACP and UCP whilst [19:00] they're not around schema.org all [19:02] provide structured data and that's how [19:04] less hallucinations happen and less [19:07] tokens are used and therefore it's more [19:09] optimal for a machine to read that [19:11] rather than something that is [19:12] unstructured. Think of it as a book with [19:14] no grammar, punctuation or even um [19:17] spelling or grammar. You know, it's [19:20] going to be much harder to read that [19:21] than it is something that's been nicely [19:23] spoonfed to them. So from a developer [19:26] mindset, this is like minimizing things [19:28] like, you know, div tags um as well, [19:31] tidying up your code, making sure that [19:33] everything's contextually interpretive [19:35] for a machine and not just a human. And [19:38] even then, uh in Google search central [19:40] in in December, we saw this slide about [19:44] product structured data. And whilst [19:46] schema.org is the glue, there are other [19:49] things such as GS1 and UNCFAC. And [19:52] depending on your business model and [19:53] vertical, you may want to look into all [19:55] the different types of structured data [19:57] that are available and adopt those [19:59] things. Now, one problem that they have [20:03] um is that structured data is only on a [20:05] page level. Um LLMs at the moment don't [20:08] always extract structured data and if [20:10] they do, they may ignore it and they [20:12] won't especially compute more complex [20:15] entity relationships. Um and this is [20:17] something that we've been thinking about [20:19] quite a lot. And then it was uh very [20:21] cool of us in November to announce that [20:24] we were partnering with the NL web team [20:26] at Microsoft. And NL web um essentially [20:30] enables AI agents to query a site's [20:33] content using natural language. So it's [20:35] like having an LLM just for your site. [20:37] It was created by RV Guha. He's the [20:40] creator of schema.org, or RSS feeds, um, [20:43] RDF, I think maybe a bit of AMP as well [20:46] and all the other web standards that you [20:48] and I love and use today. Um, and and it [20:51] was very very cool when uh, we started [20:52] working with work with RV Guhar and um, [20:55] we created a solution together. Um, [20:59] which is Yoast schema aggregation. Now [21:02] Harvey Guhar thought of a schema map and [21:04] something which I've loved all this time [21:06] but never really had the true influence [21:07] to adopt but someone like him could [21:10] think of it as an XML site map um where [21:13] it shows all the structured data of a [21:15] whole site in one endpoint and that's [21:19] all the structure under the hood. It now [21:21] provides standardized gduplicated map of [21:23] all of your entities via one single API [21:27] call. An agent no longer needs to crawl [21:29] any individual page to understand [21:31] necessarily the meaning, but it can now [21:34] ingest an entire site and therefore the [21:36] relationships between everything. And [21:37] this is just a little um a little serve [21:41] of what we have on yos.com on the pages [21:44] of how they are all listed. And it's [21:46] really really good that LLMs can [21:47] potentially look at that. Now I think [21:50] I've got a video that I can show here [21:51] and as I do that you can see how complex [21:54] a site is. This is using our schema [21:57] visualization tool um that's also [21:59] available um where you can put in the [22:02] schema map URL and then it will bring [22:04] everything in. It will show you a visual [22:06] of it and the relationship. So you can [22:07] see how complex it can get. This is [22:10] beautiful for a developer who wants to [22:12] do some diagnostics and problem solving. [22:14] It can also obviously visualize how your [22:16] site is interpreted. Not just [22:18] interpreted but it but is declared [22:21] through structured data. And it's really [22:24] not hard to in to install at all. All I [22:27] would say you do is update Yoast. It's [22:30] the in the free version and you simply [22:33] um if I go to the next slide, [22:36] simply just turn this on the schema [22:38] aggregation endpoint. And that's all you [22:40] really need to do. And what I also [22:42] thought was quite interesting is only [22:45] again two days ago, Cloudflare have [22:47] launched a slashcrawl endpoint where [22:49] this one API call can crawl an entire [22:52] site without any scripts. You can choose [22:55] between HTML, markdown like we were [22:57] talking about before or JSON. So this [23:00] data integrity and the relationships [23:02] between all of those things are going to [23:03] be so so important for an LLM to ingest [23:06] and it will create so much much more [23:09] efficiency in what's being called back [23:11] from your website. So I'd say my final [23:14] thoughts are um all the core SEO [23:17] principles still matter. Um the biggest [23:20] disruption is the reliability of the [23:22] data and not necessarily the LLMs [23:24] themselves. a lot more is going to [23:26] happen in the next year that creates a [23:28] bit more efficiency and um and a lot [23:32] less hallucinations um and a lot more [23:34] accuracy. Um I would first say audit [23:37] your content. Make sure that you've not [23:38] got any fluff. If you're one of those [23:40] companies where if I go to the about [23:41] page and talk for and and read three [23:44] sentences and I still don't know what [23:46] you do, that's for the human. That's not [23:48] for the machine. Make sure that [23:49] everything's machine readable and make [23:51] sure that everything's concise, [23:52] everything is structured well. Um, also [23:55] optimize some of your code. Your [23:56] developers might want to use a no [23:58] snippet tool alongside an SEO. Um, they [24:00] may want to optimize the div soup that [24:03] might be out there that's creating um, a [24:05] lot unnecessary code. Um, and also [24:07] ensure that all your schema entities are [24:09] connected well. Test new technologies. [24:12] So download and use an LB, test it on [24:15] your own site, see what it thinks of [24:18] your site, and then improve from there. [24:20] Um, and lastly, I would say update your [24:22] success metrics. Clicks, impressions [24:24] aren't as valuable as they used to be. [24:26] Discovery is becoming a lot more [24:28] valuable. Um, the evolution of search [24:31] isn't a risk to SEO. It's so much more [24:33] of an opportunity. And I believe that [24:36] SEOs um may become a lot more valuable [24:39] in the future to go between marketing [24:42] and product and development and be the [24:44] key and the bridge between all of those [24:46] things. Um, and yeah, that's everything. [24:49] I hope I've not been uh too over the top [24:51] or too technical or too simple, but if [24:54] we've got time for Q&A, you can ask in [24:56] here. And if not, you can connect with [24:58] me on X or LinkedIn. Thank you very [25:01] much. [25:07] Don't know if Pach is here, but if not, [25:09] feel free feel free to ask any [25:12] questions. I see three actually. Um, how [25:15] does schema markup play into the future [25:17] of SEO? Can I actually add this? I can't [25:19] add it to it. So, how does schema markup [25:22] play into the future of SEO? It seems [25:24] semi insignificant before. Um, one of [25:28] the top ways to talk to AI languages, [25:30] the way they understand. Well, at the [25:31] moment, um, some LLMs are doing, well, [25:34] all LLMs are working in a different way. [25:36] Some are taking schema, some aren't. I [25:38] know that Mark Williams Cook has done [25:40] some um experiments around it um that [25:43] say that it's giving out incorrect [25:44] information and not trusting it and not [25:46] doing enough grounding maybe. But I do [25:48] believe with things like endpoints like [25:51] the Yoast aggregation feature um and [25:54] using um the schema API as well um that [25:57] it will be able to interpret things a [25:59] lot easier with a lot more token [26:01] efficiency. Um I hope that answers your [26:04] question Jeff. Um the next one, um the O [26:08] redirection module can also create 301 [26:12] redirects for renamed or deleted pages. [26:14] Um can this be made more intelligent? [26:18] There are things being worked on. I'll [26:20] have to check with my development team [26:21] on that, but do connect with me on [26:23] LinkedIn and make sure Matthew that you [26:25] ask me that so I don't forget. Um so [26:28] yeah, I can't don't have an exact answer [26:30] for that right now. Um and taco what is [26:33] the main difference between yos schema [26:35] visualization and classy schema which [26:38] has been around for years. The answer is [26:40] is it reads the schema map. So the [26:43] entire schema map and the aggregation [26:45] tool that we've made um it crawls all of [26:48] that whereas classic schema at the [26:50] moment just crawls at a page level or at [26:53] a code level and doesn't do an entire [26:55] site in one go and visualize an entire [26:58] site like that. I hope I've answered you [27:00] there, Taco. I think that's all the [27:02] questions. How did I how did I do Peter? [27:04] Was I in time? [27:06] >> Uh, you were so uh good and in time. Uh, [27:11] I think there are actually two more [27:13] questions. Uh, would you like to give [27:15] those a stop? All right. So, uh, [27:18] Priscilla is asking, "What metrics or [27:21] ways are there to test discoverability?" [27:24] Um, I would say firstly use search [27:28] console and use web master tools. They [27:30] both have MCPs which can connect to [27:32] Claude um or whatever whatever other LLM [27:35] you're using. And as well as that, I [27:37] mean I don't want to sell our own thing [27:38] but I'm going to sell our own thing. [27:40] We've got we've got AI business insights [27:42] which you can also sign up to which [27:44] isn't anywhere near as um you know as [27:46] expensive as some of the pro tools that [27:48] may give you a bit too much information. [27:51] Um, I think it's kind of like when they [27:53] took away um ranking reports from page [27:56] two onwards, the question was, did we [27:58] even need to know what was ranking on [28:00] page 4? Was that just useless [28:02] information? Um, just just go through [28:05] everything, ask ask questions, download [28:07] and export an Excel or a CSV, put it [28:10] into an LLM and ask questions about it, [28:12] and then it will give you informed [28:14] results. [28:18] >> Okay, cool. And then we've got one more [28:21] from Matthew. This was actually the [28:23] first question asked. [28:24] >> Oh, I answered this. I answered this [28:25] one. And the answer was the answer was I [28:28] can't confirm it yet, but I've asked [28:29] Matthew to connect with me on LinkedIn [28:31] so I don't forget. And I will get you an [28:33] answer, Matthew. [28:34] >> All right, Matthew. I'm going to mark [28:36] that as done answering even if even if [28:39] that hasn't uh yet been answered. Uh [28:43] thank you, Alex. How is that for you? [28:45] >> Thank you. Oh, I'm just glad it's over [28:47] like any speaker. [28:50] I mean, and I don't look at times, so I [28:53] don't know. Maybe I that's what I think. [28:54] Am I going to be short? Am I going to be [28:56] too long? Who knows? But I don't know if [28:58] Taco says I' I'm not disappointed. He's, [29:01] uh, he's very honest with me. So, um, [29:04] I'm happy. If Taco's happy, I'm happy. [29:06] >> Uh, well, if Taco is saying you did [29:09] well, then you definitely did. Because, [29:11] as we know, Taco is very critical [29:13] towards speakers at any event. [29:15] >> Oh, he is. We worked together for a bit. [29:17] I I know. I It's just straight talking. [29:19] I call it straight talking. [29:22] >> They It's It's just the country. It is [29:24] the culture. [29:27] >> We know. Thank you so much, Alex, for [29:29] joining us today and thanks everybody [29:31] for staying Alex's session. We're going [29:33] to say uh goodbye for a little bit of uh [29:36] time and then we're moving on to the [29:38] next session where we're going to chat [29:40] about accessibility and why it is [29:43] essential in 2026. Thank you, Alex. Uh, [29:47] see you soon everybody. [29:50] >> Bye guys.  📄 **Full text:** ![[Full Document Contents/Articles/From SEO to AI Search Optimisation|From SEO to AI Search Optimisation, YouTube]] ## Highlights - what's happening. Not all the AI stuff that's happening. I know that null was talking a lot about how you can scale and improve your workflows, but I want to talk about things from a search perspective. Um, as a little bit of background, I know Petia did a wonderful introduction of me. Um, I'm principal SEO at Yoast. I'm also a co-founder and director of a small agency called Fire Cask, which is based in Manchester in the UK. Um, we do organic marketing, omni channel stuff, and as well as that, of course, um, development and design in our favorite CMS. Um, but I've actually ([View Highlight](https://read.readwise.io/read/01kkvtweb5skw7dd2ef535gerg)) - Tags: 👻 ai highlighted - on to that in a sec. Google have also released the web MCP, and it's currently there um for early preview. They've got a Discord um community as well. There's a mailing list where you can go in and see what's this about. But really, that's kind of like the new SEO for agents specifically. They go under the hood. They look at HTML and they look at elements within HTML to understand the intent of those places. So, as an example, you might use a contact form. It can identify where it is, what the ([View Highlight](https://read.readwise.io/read/01kkvtwecrds8aysjd9zs7kya6)) - Tags: 👻 ai highlighted - there's also interestingly um today that was also shared by Krishna Madavan from Microsoft. He's currently at the Friends of Search conference today in Amsterdam. um and he shared more information about the data no snippet tag which allows you to have precision content controls and it's HTML attribute where you can stay in the index and be ranked but it excludes it from snippets such as AI summaries AI overviews and AI mode. Now ([View Highlight](https://read.readwise.io/read/01kkvtwef2f4p969ehy2rzc4af)) - Tags: 👻 ai highlighted - um is that structured data is only on a page level. Um LLMs at the moment don't always extract structured data and if they do, they may ignore it and they won't especially compute more complex entity relationships. Um and this is something that we've been thinking about quite a lot. And then it was uh very cool of us in November to announce that we were partnering with the NL web team at Microsoft. And NL web um essentially ([View Highlight](https://read.readwise.io/read/01kkvtwehjdyh5xj7hhsd5693e)) - Tags: 👻 ai highlighted - enables AI agents to query a site's content using natural language. So it's like having an LLM just for your site. It was created by RV Guha. He's the creator of schema.org, or RSS feeds, um, RDF, I think maybe a bit of AMP as well and all the other web standards that you and I love and use today. Um, and and it was very very cool when uh, we started working with work with RV Guhar and um, we created a solution together. Um, which is Yoast schema aggregation. Now ([View Highlight](https://read.readwise.io/read/01kkvtwep4pr6y84ewhdrfqgb5)) - Tags: 👻 ai highlighted - much. Don't know if Pach is here, but if not, feel free feel free to ask any questions. I see three actually. Um, how does schema markup play into the future of SEO? Can I actually add this? I can't add it to it. So, how does schema markup play into the future of SEO? It seems semi insignificant before. Um, one of the top ways to talk to AI languages, ([View Highlight](https://read.readwise.io/read/01kkvtwes8ed91ymjw254gzey0)) - Tags: 👻 ai highlighted