Search Everywhere Optimization (SEO²): The 19 Surface Framework for Post Google Discovery

What Search Everywhere Optimization (SEO²) is

Search Everywhere Optimization is a framework I built to answer a question that classic SEO stopped answering well: how do you stay visible when your customers have stopped searching the way they used to? People still open Google. They also ask ChatGPT, read an AI Overview and never click a link, check Perplexity, watch a TikTok tutorial, ask Siri, tap a product photo in Google Lens, buy a thing on Amazon without ever seeing your site, scroll a Reddit thread instead of a review page, and let an agent book the appointment on their behalf. Discovery has fanned out from one search box to nineteen surfaces, and a strategy pointed at one of them leaves most of the room dark.

SEO² is what I call the practice of showing up on all nineteen with the same body of true, structured, verifiable content. The name says what it is. "Search everywhere" points at the behavior: the human act of looking something up, now happening across many surfaces at once. The superscript 2 signals succession. This is what came after the single surface era ended. It is not SEO version two the software release. It is SEO the discipline, applied to the world that replaced the one it grew up in.

The framework treats visibility as a portfolio, not a single ranking. It asks one question for every place a buyer might look: are you present, and are you trusted there? Nineteen surfaces. Nineteen answers. One method for getting to yes.

The rest of this article walks through why the shift happened, defines each of the 19 surfaces in a proper paragraph, explains the method that ties them together, and covers how to run SEO² operationally: team shape, tool stack, measurement, cadence, failure modes, and what comes next. If you already know the terrain and want the map, jump to the quick reference table at the end. If you want the whole architecture, read straight through.

SEO¹ versus SEO²: the era succession

Every generation of the web has had a dominant discovery model, and every model gets replaced by the next one. Portals gave way to search. Search gave way to social. Social gave way to a hybrid of search, social, and app store discovery. What we are living through now is the biggest handoff since Google itself: the handoff from one search surface to nineteen, and from human ranked lists to machine synthesized answers.

SEO¹ is what I call the era that ran roughly from 1998 to 2022. Google launched, blue links became the shape of discovery, and the entire craft of search optimization was built around one question: how do I rank on that page? Best practices standardized. Keywords, links, technical hygiene, on page structure, content depth, mobile responsiveness, page speed. Twenty five years of accumulated knowledge that assumed one algorithm on one surface run by one company. If you cracked Google, you had won discovery.

SEO² is what I call the era that began around 2023 and is still forming. The trigger event was public availability of large language models good enough to answer questions the way a smart friend would, and the almost immediate integration of those models into the top of the search results page. Once Google put AI Overviews above the blue links, once ChatGPT reached hundreds of millions of weekly users, once Perplexity trained buyers to expect a synthesized answer with citations, the single surface assumption stopped holding. It did not die overnight. It fractured. And into the fractures poured every other surface that had been building quietly for a decade: TikTok search, Amazon search, App Store search, Reddit search, YouTube search, voice, visual, agents. All of them, growing at once, all indexing the web on their own terms.

What did not change is the underlying craft. Clear writing still wins. Credibility still matters. Speed and structure still separate professional sites from amateur ones. What changed is where that craft has to show up, and how it gets rewarded. SEO² takes the exact skills a good SEO team already has, then applies them across eighteen more surfaces where the reward function is subtly different on each.

The temptation is to treat SEO² as a rebrand. It is not. Rebrands change labels. This is a change in the shape of the discovery graph. Once a shopper can complete a research to purchase journey without ever seeing a ten link results page, the surface that used to be the whole game is now one of many entry points. Winning it is necessary and no longer sufficient.

Why SEO² is here now: the data

Every era shift feels overdramatic when you are living through it and obvious when you look back. So let me stack the evidence for this one, because the operators I talk with underestimate how far along the shift already is.

Zero click is no longer the exception

A majority of informational queries on Google now end without a click. AI Overviews sit above the blue links, feature a synthesized answer, and answer the question well enough that most readers move on. Studies from independent trackers put the zero click share on informational queries above sixty percent, and rising on commercial ones. That is not a rounding error. That is the average reader getting what they came for and never visiting a page. If your revenue model depends on the click, you are competing for a shrinking pool.

AI answer engines have captured real query share

ChatGPT reports hundreds of millions of weekly active users. Perplexity, Claude, and Gemini each handle tens to hundreds of millions of monthly queries. A large share of those queries used to be Google searches. Even a conservative reallocation of ten to fifteen percent of research and shopping queries away from classic search is enormous in absolute terms, because the total pool is measured in trillions of queries a year.

Buyers research on social and marketplaces first

A majority of buyers under thirty five report that they start product research on TikTok, Instagram, YouTube, or a marketplace like Amazon before they open Google. That is a full generation whose discovery graph starts on a surface where classic SEO does not apply at all. Optimize for Google alone and you are invisible to them until later in the funnel, if ever.

The AI Overview click through collapse is measurable

Pages that used to rank one and pull thirty to forty percent click through can now pull single digits when an AI Overview covers the same query, even when the page is cited inside the Overview. The citation is worth something (brand mention, entity clarity, downstream trust) but the immediate traffic is smaller. Multiply that across an entire informational content library and it explains why a lot of "we still rank number one" teams are watching organic traffic slide anyway.

Agents are starting to transact

ChatGPT operator mode, Claude computer use, Perplexity shopping, and a growing set of vertical agents can now browse, compare, and complete transactions on behalf of a user. Early volumes are small. The trajectory is not. In eighteen to thirty six months, a real share of transactions in some categories will be initiated by an agent rather than a human clicking through a funnel you designed. Surfaces you have never optimized for become the ones that decide whether the agent picks you.

Retrieval augmented generation changed the citation graph

Modern AI answers are not just pulled from training data. They are assembled in real time from retrieval systems that fetch and cite sources at query time. That means fresh, well structured, crawlable content can be cited within days, not years, of publication. It also means brands with clean, indexable answer content compound faster now than in the classic SEO era, because the citation loop is measured in days rather than the six to twelve months a page used to need to earn its way into a first page ranking.

Put those together and the picture is unambiguous. Search did not shrink. It multiplied. And the operating model built for one surface cannot capture the value of nineteen without a framework that treats them as one system.

The 19 surfaces, one paragraph each

Here is the whole map. Each surface gets a proper paragraph: what it is, how it indexes, how it ranks, how it cites, and where I have gone deeper. Read straight through, or skim and come back to the ones your business touches.

Nineteen surfaces. Each with its own indexing logic, its own ranking factors, its own citation format. Almost no team has capacity to attack all nineteen at full intensity from day one, and they should not try. The point of enumerating the whole map is that when you decide what to leave for later, you leave it deliberately, not by accident.

The method: three principles that hold across all 19

Nineteen surfaces sound overwhelming until you notice the pattern underneath them. Every one of them is trying to solve the same problem: figure out which source deserves to be the answer to a question. They differ in signals, format, and speed. They do not differ in what they reward. Which means one operating method, applied faithfully, wins across all of them. SEO² reduces to three principles.

Principle 1: One source of truth per topic

Write the canonical, authoritative, verifiable answer to a topic once, publish it at a stable URL, and never fragment it across five thin pages. Fragmentation is the single biggest failure mode of the SEO¹ playbook: teams that produce ten shallow pages competing with each other for one topic while a competitor produces one thorough page that ranks, gets quoted, gets cited, and compounds. In SEO² the penalty is worse, because AI answer engines and assistants are hunting for a single source they can trust, and they will pick the deeper piece nine times out of ten. Write one master piece per real topic. Give it a permanent home. Update it in place. Everything else in the framework depends on it.

Worked example: instead of publishing "what is AEO", "AEO vs SEO", "how to do AEO", "AEO tools", "AEO examples" as five thin pages, publish one canonical AEO piece with a proper introduction, clear definitions, a comparison to SEO, a how to section, a tools section, and worked examples. That single page will outrank, outcite, and outlast the five thin pieces combined, and it becomes the citation target when other surfaces need a source for "AEO".

Principle 2: Structure it so every surface can quote it

The same source of truth has to be readable by nineteen different indexing systems, most of which cannot parse the visual layout your designer built. That means:

• Answer first. Lead every page with a direct 40 to 60 word answer in the first two or three sentences. Answer engines lift the top; readers who skim get what they came for.
• Real FAQ, real schema. Turn the questions buyers actually ask into an on page FAQ, mirrored in FAQPage structured data. Schema is how you hand a machine the answer in a format it can trust without guessing.
• Clean semantic HTML. Headings in order (H1, H2, H3), lists as lists, tables as tables, paragraphs short enough to quote. Agents and answer engines both parse the DOM, and both prefer content they can extract cleanly.
• Named authors with credentials. Every piece attributed to a real person with a real bio, real credentials, and a real presence across the web. Person schema on the byline, Organization schema on the publisher.
• Machine readable meta layer. llms.txt at the root, ai.txt where appropriate, sitemap.xml complete, structured data on every content type, Open Graph and Twitter Card metadata current.

Worked example: a well written 3,000 word canonical page on "how to choose a Shopify theme" that has a 50 word direct answer at the top, a real FAQ mirrored in schema, product comparison tables, a named author with a bio, and Article plus FAQPage plus BreadcrumbList schema on the page will rank on Google, get quoted in AI Overviews, get cited by ChatGPT and Perplexity, appear in voice results, and be picked by an agent when someone asks their assistant to compare Shopify themes for them. The same page, without the structure, ranks fine and does none of the rest.

Principle 3: Measure which surfaces cite you back, then iterate

The single biggest mistake I see teams make in SEO² is optimizing on faith. They publish, they wait, and they never look at which surfaces actually picked their content up. Meanwhile a competitor with worse content but better measurement is iterating twice as fast. The fix is a citation measurement stack that covers all nineteen surfaces, run on a fixed cadence.

For each priority topic, keep a small tracking prompt set (10 to 25 questions a real buyer would ask) and check every two weeks: does your content appear in Google AI Overviews, Perplexity, ChatGPT, Claude, Gemini, TikTok search, YouTube search, Amazon (if relevant), and the local pack (if relevant)? Log it. Do not rely on gut feel. The surfaces that never cite you tell you where to fix the source of truth. The surfaces that do cite you tell you what to repeat.

Worked example: after six weeks of tracking, you notice that Perplexity and Google AI Overviews both cite your canonical AEO page, but ChatGPT and Claude keep citing a competitor. The gap is almost always one of three things: your entity signals are weaker (Wikipedia entry, Wikidata entry, third party mentions), your citation graph is thinner (backlinks and Reddit mentions), or the competitor has original data you do not. Address whichever gap explains the miss and reprompt in three weeks.

These three principles are not opinions. They are the operating system. Every other tactic in SEO² is a specialization of one of these three moves applied to a specific surface. The full articulation of principle one and principle two lives in The Structured Truth Method, which is the SEO² operating system in shorter form.

What ties the 19 surfaces together

The reason SEO² is a framework and not a checklist of nineteen unrelated jobs is that the surfaces share more foundation than they signal. Every one of them reads a version of the same underlying primitives, and getting those primitives right pays out across all of them. Four foundations do most of the work.

Structured content

Answer first prose, real headings, real FAQ, real tables, real schema. Every surface reads structure. Classic search reads it for ranking. Answer engines read it for extraction. Assistants read it for citation. Agents read it to complete tasks. Voice reads it for speaking. Visual reads it via alt text and product data. Structured content is the closest thing to a universal input any of the nineteen surfaces will accept.

Credibility signals

Named authors with real credentials, consistent bios across the web, third party mentions in credible outlets, backlinks from reputable domains, on page transparency about who wrote what and when. Every surface weighs credibility, and the weight is going up, not down, as generative systems try to avoid citing untrustworthy sources. E-E-A-T started as a Google shorthand and has become the shared trust language of the whole stack.

Entity clarity

A brand or person is an entity, and every surface reads entities to know who is who. Consistent name, logo, description, category, links, dates, affiliations across your own site, Wikipedia, Wikidata, LinkedIn, Google Business Profile, industry directories, and third party mentions. Entity clarity is what makes an AI assistant confident enough to name you unprompted. It is what makes a voice system pick you as the source. It is what makes a knowledge panel show up correctly. Entity work is the invisible backbone of SEO².

Freshness discipline

Every surface prefers current content over stale content. Not because "fresh" is a value in itself, but because current signals accuracy, and accuracy is the closest thing to a universal ranking factor. Update your canonical pages in place, timestamp them, republish when the underlying facts change, retire pages that will not be maintained.

Where the nineteen surfaces separate is in the specifics of how each one indexes, ranks, and cites. Voice cares about spoken language phrasing in a way Amazon does not. Amazon cares about star rating and prime eligibility in a way voice does not. TikTok cares about audio transcripts and cover frames in a way Google organic does not. But those specifics are the last mile. Nail the four foundations first and the last mile becomes a series of small adaptations rather than nineteen full time strategies.

How to run SEO² operationally

A framework you cannot operate is theater. Here is what running SEO² actually looks like as a team, a tool stack, a cadence, and a roadmap. This is the same shape I use when I set the discipline up inside an operating company.

Team shape

You do not need nineteen specialists. You need one or two generalists who understand the framework, plus specialist support in the two or three surfaces that matter most for your revenue. For most operators the shape looks like:

• One SEO² lead. Owns the framework, the source of truth library, the roadmap, and the measurement stack. Reports directly to the head of marketing or the founder.
• One content producer. Writes the canonical source of truth pieces, works with subject matter experts, maintains the update cadence.
• One technical partner. Owns crawlability, schema, page speed, and the machine readable meta layer. Can be internal or a specialist agency retainer.
• Two or three surface specialists on retainer. For most ecommerce it is MSO (Amazon), PFO (feeds), and SSO (TikTok or Instagram). For most local service it is LSO and CSO. Retainer, not full time, until volume justifies dedicated staff.

That is four to six people, most of them part time on this scope, running a discipline that used to require twelve. The efficiency comes from the source of truth model: one canonical piece serves nineteen surfaces, so you are not staffing nineteen separate content teams.

Tool stack

Every operator picks their own, but a workable SEO² stack looks like this:

• Classic SEO: Google Search Console, GA4, Semrush or Ahrefs, Lumar or Screaming Frog for crawl audits.
• AEO and GEO citation tracking: Profound, Scrunch, Semrush AI Overviews, BrightEdge AEO, or Otterly. Also manual weekly prompt tests logged in a spreadsheet.
• Schema and technical: Google Rich Results Test, Schema.org validator, Lighthouse for CWV, PageSpeed Insights.
• Marketplaces (MSO): Helium 10, Jungle Scout, Amazon Brand Analytics, Walmart Retail Link.
• Local (LSO): BrightLocal or Whitespark, Google Business Profile Insights, Semrush Local.
• Social and video (SSO, VDO): TikTok Creative Center, Instagram Insights, YouTube Studio, VidIQ or TubeBuddy.
• App store (ASO): App Store Connect, Google Play Console, Sensor Tower, App Radar.
• Feeds (PFO): Google Merchant Center, DataFeedWatch, GoDataFeed, Channable.
• Entity and knowledge (KGO): Kalicube, Diffbot, Wikidata, Google Knowledge Graph API.

The failure mode with tooling is buying too much too soon. Start with Search Console plus GA4 plus one AI visibility tool plus a spreadsheet of prompt tests. Add per surface tools as you prove the surface is worth the retainer.

Cadence

Discipline is what separates SEO² teams that compound from SEO² teams that publish and hope. A workable cadence:

• Weekly (30 minutes): Prompt test the 10 to 25 tracking questions across ChatGPT, Claude, Perplexity, Gemini, and Google AI Overviews. Log the citations.
• Biweekly (2 hours): Review Search Console for classic SEO, review AI visibility tool for citation share, review top three surfaces you are actively working. Assign the next two weeks of edits and publishes.
• Monthly (half day): Full source of truth library audit: what needs updating, what needs republishing, what needs a new canonical piece. Refresh timestamps.
• Quarterly (full day): Roadmap review across all 19 surfaces. Which surfaces are we deliberately leaving? Which are we adding? What is the next canonical piece we owe the library?
• Annually: Full entity audit (Wikipedia, Wikidata, Google Knowledge Graph, LinkedIn, industry directories). Reconcile facts across the web.

Roadmap sequencing

Nobody attacks nineteen surfaces at once. The right order for most operators:

1. Months 1 to 2: Foundation. Fix technical SEO, publish llms.txt and ai.txt, ship structured data on every content type, clean up entity facts across the web.
2. Months 2 to 4: Source of truth library. Publish or consolidate the ten most important canonical pieces for your business. Rewrite for answer first, add real FAQs, apply schema.
3. Months 3 to 6: The two surfaces past classic search that matter most for your revenue. For most that is AEO plus one of (LSO, MSO, SSO).
4. Months 6 to 12: GEO. Original data pieces. Reddit and community presence. Author entity work.
5. Year 2: Add remaining surfaces in the order that maps to buyer behavior in your category.

Every operator's exact sequence differs, but the shape holds: foundation first, source of truth library second, two or three surfaces third, then expand. Trying to hit all nineteen in the first quarter is how teams burn out and see no compounding.

A worked example: SEO² for one topic across all 19 surfaces

Frameworks are easier to trust when you can see one work end to end on a single topic. So let me walk through what SEO² looks like applied to one canonical topic for a hypothetical company. Pick something concrete: a mid market ecommerce brand that sells premium reusable water bottles, and their canonical topic is "how to choose a reusable water bottle". Here is what publishing that piece under SEO² discipline looks like across every one of the 19 surfaces.

Foundation (do this before anything else)

Before the piece ships, the brand's technical stack has to be right. Site loads under 2.5 seconds LCP on mobile, INP under 200 ms, CLS near zero. Structured data on every product and content page. llms.txt at the root, ai.txt where relevant. Google Business Profile filled out if there is a physical or service location. Wikipedia and Wikidata entries for the brand if it qualifies, and consistent brand facts (name, description, category, founding year, logo, product line) across LinkedIn, Instagram, TikTok, Amazon storefront, industry directories, and any third party mentions.

Without this foundation, nothing else scales. With it, the canonical piece becomes an amplifier.

The canonical source of truth (Principle 1 in action)

One page. One URL. Roughly 3,500 words. Titled "How to Choose a Reusable Water Bottle: The Complete Guide". Answer first: a 55 word direct answer at the top that tells a reader (or an answer engine) the four things that actually matter (material, capacity, insulation, closure type) and the tradeoffs among them. Then a proper piece: material deep dive (stainless steel vs glass vs Tritan vs silicone), capacity guidance by use case, insulation performance by test data, closure comparisons, care instructions, common mistakes, a comparison table, a real FAQ with 12 questions buyers actually ask, and a bio for the named author who wrote it.

Schema: Article, FAQPage, BreadcrumbList, Person (author), Organization (publisher), and Product references to the brand's own SKUs where relevant. Speakable schema on the direct answer and the FAQ.

How the canonical page plays across all 19 surfaces

Now watch the same page compound across the framework.

1. SEO: ranks for "how to choose a reusable water bottle" and forty related long tail queries. Internal linking pulls authority from category and product pages. Backlink outreach targets the ten most credible third party sources that would cite a guide like this.
2. AEO: the 55 word direct answer at the top gets extracted into Google AI Overviews and Perplexity for the target query and its variants. The FAQ schema fills in the "People also ask" boxes on classic Google results.
3. GEO: ChatGPT, Claude, and Gemini learn to associate the brand with the topic through repeated retrieval of the canonical page plus third party citations of it. Within twelve weeks, the brand starts appearing by name in assistant answers on the topic.
4. AAO: the product SKUs referenced in the guide use structured product data, clean URLs, and standard cart mechanics, so an agent booking a "buy me a reusable water bottle" task can complete the transaction without heroics.
5. VSO: the direct answer, phrased for a human reader, also reads naturally when a voice assistant speaks it. Siri, Google Assistant, and Alexa pick up the answer for spoken queries.
6. VxSO: every product image in the guide uses descriptive alt text, structured product data, and consistent naming, so Google Lens and Pinterest Lens surface the brand's SKUs when a shopper snaps a bottle in a coffee shop.
7. ASO: not directly applicable to a content page, but if the brand ships a companion app for hydration tracking, the app store listing borrows the same category and keyword research the guide surfaced.
8. KGO: the guide reinforces the brand's entity in Google's Knowledge Graph as an authority on hydration and reusable containers. Consistent Person schema on the author reinforces author entity too.
9. LSO: if the brand has retail locations, the guide is linked from Google Business Profile posts and used as anchor content for local link building with regional press.
10. CWV: the page passes Core Web Vitals cleanly, which means every other surface trusts it more. AI answer engines and agents both prefer fast sources.
11. E-E-A-T: named author with real credentials, on page bio linking to LinkedIn and other credible profiles, third party mentions of the author across the web, dates on the page, sources cited inline.
12. GLOBO: if the brand ships internationally, the guide is translated (not just machine converted) into the top three locale variants with hreflang tags and locale specific product data.
13. Web3: not primary here, but the brand maintains a stable identity presence in emerging identity registries so that agent commerce infrastructure can resolve them cleanly.
14. MSO: the same brand messaging, category positioning, and comparison framing used in the guide is applied to the brand's Amazon and Walmart listings, so the marketplace SEO reinforces the same story.
15. SSO: the guide is repurposed as a series of TikTok and Instagram videos: one on material tradeoffs, one on capacity by use case, one on insulation testing, one on closure types. Captions target the same queries the guide targets. Videos link back to the guide.
16. VDO: a longer form YouTube version of the guide gets published with chapter markers matching the guide's sections, and transcripts filed for search extraction.
17. PFO: product feeds into Google Merchant Center and Meta Commerce use the same taxonomy and attribute language the guide validated. Shopping ads and organic shopping surfaces reinforce the guide's authority signals.
18. FEO: the guide, or seasonal variants of it (hydration for summer, insulated bottles for winter), get pushed into Google Discover through the fresh, categorically relevant, high engagement content the feed prefers.
19. CSO: the brand's authors participate in Reddit and Quora threads on hydration and reusable containers as genuine helpful commenters (not brand shills), and the guide gets cited organically. Those citations feed into AI answer engines when AI systems mine Reddit for real experience answers.

One canonical page. Nineteen surfaces. Same body of structured truth, adapted for each surface's format, indexing logic, and citation mechanics. The work of writing the piece is a fraction of what a team used to spend producing nineteen separate content strategies, and the compounding is stronger, because everything reinforces everything else.

That is what SEO² looks like when it works. And it is why fragmentation (nineteen shallow pieces for nineteen surfaces) actively costs you: the compounding stops.

The signals that win each surface, side by side

Every surface has its own dominant ranking signal, and knowing the shape of it is the difference between publishing content that ranks and publishing content that ranks, cites, and converts. The pattern I use to hold this in my head:

Every one of these signals traces back to the four foundations (structured content, credibility signals, entity clarity, freshness discipline). What varies is the specific mechanism each surface uses to read those foundations. Get the foundations right and the surface specific tuning becomes a matter of format, not a matter of restarting.

Why SEO² compounds harder than SEO¹ did

Classic SEO compounded, but slowly and unevenly. A page took six to twelve months to rank, backlinks accumulated at whatever rate outreach could produce, and the compounding curve on any single piece plateaued once it reached its natural ceiling on the target keyword. SEO² compounds faster and with a steeper curve for four reasons, and understanding them is why the early movers get outsized returns.

AI training data has a memory

Every AI model that indexes the web builds a persistent representation of who is authoritative on what. Once a model "learns" that your brand is the source for a topic (through repeated exposure across training data plus retrieval plus citation graphs), it keeps naming you until something displaces the association. That association is sticky in a way a Google ranking never was. A Google ranking can be lost in an algorithm update overnight. A model level entity association takes months of countervailing signal to shift.

Citation graphs reinforce themselves

Every time you get cited by an AI answer, that citation becomes a signal to other AI systems that you are worth citing. Being quoted by Perplexity feeds into the training and retrieval systems of the next model. Being named in ChatGPT gets screenshotted, tweeted, and turned into new content that itself feeds back into the graph. This is a feedback loop that runs faster than the backlink graph ever did, because the units are prose citations rather than markup, and they multiply through the same synthesis systems that consume them.

Entity clarity accumulates

Every consistent brand and person fact you publish across the web (Wikipedia, Wikidata, LinkedIn, Google Business Profile, industry directories, third party mentions) reduces the ambiguity every surface has about who you are. Reduced ambiguity means higher confidence, which means more mentions, which means more entity signals, which reduces ambiguity further. This is compounding at the entity layer that did not exist as a first class asset in SEO¹.

First mover advantage in AI answers

The number of brands actively optimizing for AEO and GEO is still small. The number optimizing for AAO is smaller. Which means the competitive landscape on the newest surfaces is thinner than classic SEO has been for a decade. Early presence gets baked into model behavior in a way that will be expensive to displace once the space is crowded. This will not last. The window is measured in quarters, not years, and it is closing fastest in the categories where AI answer engines are already the primary discovery surface (finance, health, travel, tech, consumer goods).

The math of compounding matters here. A team that starts SEO² today and compounds for eighteen months will be structurally ahead of a competitor who starts eighteen months from now, even if the competitor spends more, because entity clarity and citation graph share are not proportional to spend. They are proportional to time in market plus discipline. Time you cannot buy back.

Common failure modes

The failure modes in SEO² are consistent enough across operators that I can list them by frequency. If you are running the discipline and something feels off, work through this list first.

Treating AI Overviews as regular SERPs

The most common failure is a team that keeps optimizing pages for the classic ten link results page while an AI Overview sits above them capturing sixty percent of the query. They still "rank number one" on the queries they track, and organic traffic keeps sliding, and nobody connects the two. The fix is to add AI Overview citation as a first class metric alongside classic ranking, and to rewrite pages so they are extractable (answer first, schema, FAQ) rather than just rankable.

Blocking AI crawlers reflexively without a strategy

The second most common failure is a team that read a headline about AI companies training on scraped content and rushed to block GPTBot, ClaudeBot, PerplexityBot, and Google Extended across the whole site. Now their content is invisible in the answers those systems generate. The fix is a considered access policy: default open for public brand and product content that you want cited, selectively blocked for paywalled archives and subscriber only material. Reflexive blocking is the fastest way to be invisible in SEO².

Publishing content that ranks but never gets cited

You have a page ranking one on the target keyword. It gets zero AI citations. This is almost always a structure problem, not a content problem. The page is well written for a reader but not extractable by a machine: no direct answer at the top, no real FAQ, no schema, headings out of order, walls of prose without lists or tables. Rewrite the top of the page to answer first in 40 to 60 words, add a real FAQ mirrored in FAQPage schema, and put the extractable facts in structured lists and tables. Citation follows within weeks.

Optimizing for keyword ranking while the buyer is asking questions

Classic SEO keyword research still yields high volume keywords. Those keywords are increasingly not what buyers actually type into ChatGPT or ask their assistant. They ask longer, more conversational, more specific questions. If your content only targets the two word keyword, you win the ranking and lose the answer. Extend keyword research to include the question forms buyers use with AI assistants (which is where the AI visibility tools earn their keep) and write pages that answer those question forms directly.

Fragmenting the source of truth across shallow pages

The classic content marketing playbook produced volume: ten pages per topic, each targeting one keyword variant. In SEO² this is actively counterproductive. AI systems want one deep, authoritative source, and they will pick the deeper piece over the shallow one every time. Consolidate fragmented pages into canonical master pieces per real topic. Redirect the shallow variants to the canonical piece. Depth beats breadth on every one of the nineteen surfaces.

Skipping entity work

A team publishes great content and cannot understand why AI assistants still cite the competitor. The gap is usually entity clarity. The competitor has a Wikipedia article, a Wikidata entry, a filled out Google Knowledge Panel, consistent bios across the web, and a named author who is a recognized entity in the space. You have none of that. Entity work is invisible in dashboards and enormous in impact. It is the single highest leverage lever most SEO² programs are missing.

Treating the framework as a rebrand instead of a discipline

The failure of intent: a marketing team hears "SEO²", updates the pitch deck, adds three bullets to the strategy document, and keeps doing exactly what they did before. The discipline is the whole point. Without cadence, measurement, and the source of truth library, the label is just noise. The teams that get outsized returns are the ones that treat SEO² as an operating change, not a positioning change.

What comes next in SEO²

SEO² is not the last shape discovery will take. It is the shape that fits the surfaces that exist today, and the framework is designed to absorb new surfaces as they emerge. Three things are worth watching over the next twelve to twenty four months.

AAO will become as important as AEO and GEO

Agent Action Optimization is where the next layer of value gets captured. Once a real share of transactions in a category get initiated by an agent (an assistant that browses and buys for a human), the question moves from "does the agent know about me" (GEO) to "does the agent pick me" (AAO). Winning AAO requires that your site be agent readable: clean semantic HTML, deterministic forms, transparent pricing, structured product data, working accessibility, and clear affordances for programmatic completion of common tasks. Most sites are not close to ready. The ones that get ready in the next twelve months will be structurally advantaged as agents scale.

llms.txt will formalize

The llms.txt spec is emerging as the AI world's equivalent of robots.txt: a plain text file at the root of your domain that tells AI systems what your site is authoritative on and where the canonical facts live. Today it is a convention, not a standard, and adoption is uneven. Within twelve months it will almost certainly firm up into a widely followed spec, and brands that already publish one will be treated as first class citizens by the AI stack. The cost of publishing one is nearly zero. There is no reason to wait. Guide: The llms.txt Guide.

Search will disaggregate into a dozen query modes

The trajectory is not "AI replaces search". It is "search fragments into query modes that each get their own optimal surface". Informational questions go to AI Overviews and assistants. Comparison shopping goes to marketplaces and PFO surfaces. Product research goes to social search and video. Local intent goes to map surfaces and LSO. Complex research goes to Perplexity and long form assistants. Voice queries stay with voice assistants. Each mode will get more specialized, and the sites that can serve all of them from one body of structured truth will win the disaggregated share.

SEO² is deliberately built to absorb this. When the twentieth surface emerges (and it will), the framework does not break. It gets a new surface entry, a new set of specialist tactics, and the same underlying source of truth continues to serve. The point of a framework is that it survives the arrival of what it did not anticipate.

Quick reference: all 19 surfaces

The whole map in a scannable format. Bookmark this table.

Who created SEO²

I am Frederick Sona. I named and defined this framework. I have spent fifteen plus years as a full stack growth leader across technology, advertising, marketing, design, and sales, and ten of those years as CMO and Creative Director at Inkgility. I started as a programmer writing code as a kid, and I still ship the systems I write about; the site you are reading is one of them. Over the years I have run a client's largest Google Ads budget at roughly three hundred thousand dollars a month, grown a local business from around forty reviews to more than two hundred inside a year, and built the content and technical systems behind ecommerce operations across print, digital, and hybrid categories. SEO² is the framework I use to keep the brands I work with visible as discovery moves off the classic search page and into the nineteen surfaces that replaced it. If you want the shorter operating manual behind SEO², read The Structured Truth Method. If you want to talk about running the framework in your company, come say hello at fredericksona.com.