Modern SEO: Optimizing for search engines and LLMs

Search has changed. Learn how SEO, AI, and credibility now shape discoverability in modern search and LLM-driven results.

Transcript

1. Grant Simmons LI: /in/simmonet X: /simmonet W: grantsimmons.com
2. Modern SEO: Optimizing for search engines and LLMs How AI search builds on classic SEO and information retrieval concepts (and what we can do about it!) C Content C Code E Entities E Evidence R Retrieval Clarity · Relevance · Evidence
3. Zach talked about how you earn credibility for your content in the world. I’m talking about how you make that credibility legible to retrieval systems. For that I’m going to rely on one of the worlds most infamous detectives...
4. Zach talked about how you earn credibility for your content in the world. I’m talking about how you make that credibility legible to retrieval systems. For that I’m going to rely on one of the worlds most famous detectives...
5. And thanks (Simpsons) Kent!
6. Let’s Get This Out of the Way!
7. AI search did not ‘kill’ SEO It made the old fundamentals easier to see. AI answers still depend on discovery, interpretation, relevance, evidence, and trust. The interface changed. The retrieval problem did not. SEO IR AI Search Search practice Retrieval science
8. The progression we need to make Classic SEO Help engines find and understand pages Information Retrieval Match an information need to evidence Entities Clarify the thing being discussed Schema Make meaning explicit in code Entity Maps Package knowledge for non-human bots AI Search Retrieve, synthesize, cite, answer The goal is not “new SEO.” The goal is clearer retrieval. AI Search × Classic SEO × Information Retrieval
9. Classic SEO solved a retrieval pipeline 1 Discover Can the system find it? 2 Crawl Can it access it? 3 Understand Can it interpret it? 4 Select Can it match the need? 5 Present Can it show the useful result? AI search changes how results are assembled. Not the need for this pipeline. Nor the end-user (a human).
10. Information retrieval is the underlying game Information Need Representation Retrieval & Ranking Evidence
11. Information retrieval is the underlying game Search is a representation problem before it is a ranking problem. Better representation makes relevance easier to compute and evidence easier to trust Information Need Representation Retrieval & Ranking Evidence
12. A useful credo for AI search “Information retrieval rewards clarity, relevance, and evidence. LLMs make those signals more visible, not less important. Clarity Relevance Evidence
13. Traditional SEO still maps to retrieval signals Accessible corpus Document representation Relationship graph Explicit meaning Evidence & credibility AI optimization works best when it strengthens these same signals. AI Search × Classic SEO × Information Retrieval Crawlable architecture Titles, headings, copy Internal links Schema markup Digital PR / citations
14. Ambiguity is the enemy of retrieval The same word can point to different things. A retrieval system has to decide which thing the content means before it can select evidence. Animal Car brand Sports team Same string ≠ same thing Jaguar
15. Entities provide clarity An entity is the thing your content is about. People, organizations, products, places, concepts, events, and metrics become anchors for retrieval. S String T Thing R Relation E Evidence A Answer AI Search × Classic SEO × Information Retrieval
16. Keywords are surface forms; entities are anchors AI share of voice LLM visibility share Generative search visibility Canonical entity “AI Share of Voice” Clear evidence for retrieval Our job is to collapse messy language into stable meaning. AI Search × Classic SEO × Information Retrieval AI SOV
17. Knowledge graphs make relationships explicit authored by supported by sameAs Relationships make meaning computable AI Search × Classic SEO × Information Retrieval Google described the Knowledge Graph as modeling real-world entities and relationships — “things, not strings.” described by Author Evidence Source Page Topic Entity
18. Schema v1: “How do we get rich results?” The familiar reason to implement schema: eligibility for enhanced search features Useful - but incomplete as the main mental model AI Search × Classic SEO × Information Retrieval Google Search Central: structured data can enable rich results and helps Google understand page content. Rich result ratings price availability
19. Schema v2: “How do we clarify meaning?” Schema becomes a clarification layer. Not just markup for display. Markup for interpretation. Product page @type name description sameAs Machine can resolve what this page means Rich results are a possible output. Clarification is the strategic input.
20. What schema clarifies { "@type": "Organization", "@id": "#org", "name": "…", "sameAs": ["…"], "knowsAbout": ["…"] } 1 Type What kind of thing is it? 2 Identity Which thing is it? 3 Attributes What facts describe it? 4 Relations How does it connect? 5 Evidence Where is it supported?
21. SameAs anchors identity SameAs says: “this is the same thing as that known thing.” It helps a system disambiguate your entity against stable reference points. Your entity @id: #org sameAs Reference URLs Wikidata · Wikipedia · official profiles Open knowledge graph SameAs is not decoration. It is an identity bridge. AI Search × Classic SEO × Information Retrieval Schema.org defines sameAs as a URL of a reference page that unambiguously indicates identity.
22. Internal links express relationships Internal links are not just navigation. They show which entities and topics explain, support, compare, or depend on each other. Core topic Definition Example Evidence defines applies supports proves Anchor text is a relationship label.
23. Important entities need canonical homes Give every core entity a place to resolve. A canonical entity home defines the thing, names variants, links supporting evidence, and points outward with SameAs. Entity Home Definition Variants Evidence SameAs No home = more inference. More inference = more ambiguity. HERE’S NO PLACE LIKE HOME
24. Triangulation creates confidence Machines trust repeated, consistent signals. Schema, SameAs, and internal links should point to the same entity story. Schema SameAs Links Entity clarity The stronger the triangulation, the less the model has to guess.
25. sitemap.xml Tells crawlers what pages exist From sitemap to entity map entitymap.json Tells agents what knowledge exists Pages are the container. Entities are the meaning. Entitymap.org
26. entitymap.org: an open standard for AI-readable content EntityMap publishes a structured, entity-first index of website content. Designed for AI agents, LLMs, and retrieval-augmented systems - a discovery layer for site knowledge. Stable v1.0 entity-first machine-readable /entitymap.json entities[] relations[] chunks[] NOT: llms.txt entitymap.org
27. What an EntityMap gives a bot Entity name, type, description Identity sameAs, canonical label Relations typed connections Evidence source chunks + URLs Publisher attribution + freshness Less guesswork for non-human bots This is clarity as infrastructure.
28. Entity maps help reduce retrieval loss Disambiguation loss Many names for one thing get split apart Attribution loss Evidence is used without clear publisher identity Reasoning loss Relationships stay buried in prose Entity-first evidence layer
29. Fundamentity is the wrapper Fundamentity ties: “meaning for retrieval” to “content for ranking.” It is a way to keep AI search strategy grounded in classic SEO while making entities explicit. Content for ranking Fundamentity Meaning for retrieval The wrapper keeps entity work from becoming disconnected from SEO outcomes.
30. Content ranks; entities retrieve Content for ranking usefulness depth coverage quality links Entities for retrieval identity relationships evidence source clarity sameAs Fundamentity The strategy is the bridge, not either side alone. AI Search × Classic SEO × Information Retrieval
31. Fundamentity operating model 4. Retrieval layer AI search, answer engines, RAG, citations 3. Evidence layer chunks, claims, URLs, publisher attribution 2. Entity layer schema, SameAs, internal links, EntityMap 1. Content layer useful pages, topical depth, experience, UX Fundamentity wraps the stack
32. Implementation path: build clarity in layers 1 Inventory core entities 2 Assign canonical homes 3 Clean internal links 4 Add schema as clarification 5 Anchor identity with SameAs 6 Publish entity map 7 Validate and measure Do not start with markup. Start with the entities you want machines to understand.
33. Measure clarity, not just implementation Ask whether a machine can answer 4 questions Who / what is this? What is it related to? Why is it relevant? Where is the evidence? Entity coverage Schema consistency SameAs quality Orphan entity count AI answer inclusion Citation accuracy Measure the following:
34. Final takeaways Schema clarifies. Think beyond rich results. Links contextualize. Internal links express relationships. Entity maps package evidence. For bots that need meaning, not navigation. Entities provide clarity. Clarity makes retrieval easier. Retrieval shapes AI visibility.
35. Two final nuggets of wisdom (you can quote me)
36. Schema is not the content. Schema is the clarification layer around the content. Bad schema does not make weak content strong. Good schema makes real content easier to understand.
37. Schema is not the content. Schema is the clarification layer around the content. Bad schema does not make weak content strong. Good schema makes good content easier to understand.
38. Entities and an Entity Map are not magic AI visibility tools. They are a way to make context, relationships, and evidence easier for non-human systems to inspect. But don’t forget the humans ☺
39. Entities and an Entity Map are not magic AI visibility tools. They are a way to make context, relationships, and evidence easier for non-human systems to inspect (and understand) But don’t forget the humans ☺
40. Entities and an Entity Map are not magic AI visibility tools. They are a way to make context, relationships, and evidence easier for non-human systems to inspect (and understand) But... don’t forget the humans ☺
41. Entities and an Entity Map are not magic AI visibility tools. They are a way to make context, relationships, and evidence easier for non-human systems to inspect (and understand) But... don’t forget the humans ☺
42. Grant Simmons LI: /in/simmonet X: /simmonet Search Google for: “Grant Simmons SEO”
43. Sources and reference points Google Search Central SEO Starter Guide: SEO helps search engines understand content and helps users find sites. Google Search Central Structured Data intro: structured data gives explicit clues about page meaning and can enable rich results. Schema.org sameAs: a URL that unambiguously indicates an item’s identity. Google Blog Knowledge Graph: entities and relationships; “things, not strings.” EntityMap.org EntityMap specification v1.0: structured, entity-first index for AI agents, LLMs, and RAG pipelines. Stanford IR book Information retrieval foundations: representation, relevance, ranking, and evidence. Start with the entities you want to be known for. Then make them unambiguous. InLinks.com / Waikay.io SEO / GEO Entity Tools to help do everything I talked about today – Maybe not *everything*! ☺