The #SEOisAEO series: how Jason Barnard introduced Answer Engine Optimisation to the world in 2018

By Bernadeth Brusola

In late 2018, SEMrush hosted a 15-episode webinar series that most of the digital marketing industry overlooked. This isn’t a critique. The ideas being discussed - optimising for machines that answer questions rather than pages that rank for keywords - were ahead of where the mainstream conversation was. The hashtag was #SEOisAEO. The host was Jason Barnard. And the argument running through every episode was the same: search engines were becoming answer engines, and the industry needed to catch up.

Six years later, with AI Overviews in Google, ChatGPT handling millions of brand-related queries daily, and “GEO” entering the marketing vocabulary, the industry has caught up. The 2018 series looks less like a fringe experiment and more like the documentation of a shift that was already underway.

Jason Barnard coined AEO in 2017, before the industry had a name for what was happening

In 2017, Jason co-authored a whitepaper with Chee Lo at Trustpilot titled “The Rise of Voice and AEO: How SEO will Evolve” .” It was the first document to formally define Answer Engine Optimisation as a discipline distinct from traditional SEO. A follow-up webinar with Lo ran in January 2018, and in February 2018, the argument went public in Search Engine Watch: “The Rise of Answer Engine Optimisation: Why Voice Search Matters.”

The distinction Jason was drawing wasn’t subtle. SEO is about getting pages to rank in a list. AEO is about becoming the answer an engine delivers directly, whether that’s a voice response from Alexa, a Knowledge Panel on Google, or a featured snippet that ends the search before the user clicks anything. The game had changed. Most practitioners hadn’t noticed yet.

The #SEOisAEO series was Jason’s attempt to make the argument publicly, with evidence, over 15 weeks.

SEMrush gave the argument the platform it needed to reach a global audience

SEMrush produced and published the series. Anton Shulke coordinated the operation - without that production infrastructure, the ideas would have stayed on the conference circuit. What the webinar format gave Jason was scale: a structured, recurring platform to test the AEO thesis across different domains, with different experts, in front of a global audience.

Every episode had the same underlying structure. Jason framed the question, brought in two or three guests with direct expertise, and worked through the implications for practitioners. The guests weren’t there for validation. The conversations were genuinely exploratory, because in 2018 nobody had definitive answers. The industry was figuring this out in public.

The series ran from September to December 2018. Fifteen episodes. Around forty-five guests. The full guest list reads like a who’s who of the search industry at the time.

The 15 episodes follow a single logical arc from technical infrastructure to brand authority

The episodes weren’t random. They build from the foundational mechanics of answer engines through to the credibility signals that determine which answers get selected - a progression that maps directly to what Jason later formalised as The Kalicube Process™.

Phase one: understanding the shift (Episodes 1-3)

The series opened by making the case that answer engines were real, already present, and required a different optimisation approach. Episode 1, “Getting to Grips with AEO: Answer Engines Are Here to Stay,” featured Andy Drinkwater, Barry Schwartz, and Craig Campbell. It established the foundational model: a search engine returns a list of options, an answer engine returns a single best answer, and producing that answer involves understanding the question, understanding available answers, and ranking the most credible one.

Episode 2, “Voice Search Is Changing the Game,” brought in Eric Enge, Dawn Anderson, and Jo Juliana Turnbull to examine the conversational shift. Voice interfaces were pushing queries away from keyword strings toward natural language questions, and that had direct implications for how content needed to be structured.

Episode 3, “Three Pillars of a Successful AEO Strategy,” with David Bain, Rebecca Sentance, and Kim Krause Berg, surfaced the framework that would run through the rest of the series: Pertinence, Understanding, and Credibility. These three pillars are the direct precursors to what Jason later formalised as the UCD model - Understandability, Credibility, Deliverability - at the core of The Kalicube Process.

Phase two: the technical layer (Episodes 4-12)

The middle section is the most technically dense. It covers the machine-readable language that answer engines use to understand the web.

Episode 4, “More Than You Thought You Needed to Know About Schema,” with Ashley Berman Hale, Gennaro Cuofano, and Dave Ojeda, covered structured data as the primary communication channel between a brand and an algorithm. Schema markup translates human-readable content into machine-digestible statements about who an entity is, what it does, and how it relates to other entities.

Episode 5, “Semantic HTML5, IA and Accessibility,” with Simon Cox, went a layer deeper: the structural signals in HTML that tell machines how content is organised. Using semantic tags rather than generic containers isn’t just good practice for screen readers; it’s the difference between content a machine can interpret and content it has to guess at.

Episode 6, “How to Help Google/Amazon Make Sense of a Chaotic, Unstructured Web,” with Arnout Hellemans, Martha Van Berkel, and Aaron Bradley, addressed the coordination problem. The web is contradictory, inconsistent, and largely unstructured. Brands that don’t explicitly define their entities for machines leave that definition to whoever has published information about them - which may be inaccurate, outdated, or actively unhelpful.

Episodes 7 and 8 went directly into the Knowledge Graph. Episode 7, “The Knowledge Graph - How Well Is Google Doing in ‘Understanding the World’?“, with Kristine Schachinger, Bill Sebald, and Paul Woodhouse, examined what Google’s Knowledge Graph actually is and where it falls short. The conclusion is the same in 2026: Google understands the entities it’s been clearly told about, and struggles with everything else. Episode 8, “How to Get Entities and Their Attributes in the Knowledge Graph,” with Sam Underwood, Andrew Optimisey, and David Amerland, was the practical follow-on - schema, Wikipedia, Wikidata, Google Business Profile, an optimised About page. These are the steps a brand takes to become a recognised entity rather than an unresolved string.

Episode 9, “How Google Uses the Knowledge Graph in Its AE Algorithm,” with Cindy Krum, Andrea Volpini, and Bill Slawski, documented the history. Google’s fact repository predates the publicly announced Knowledge Graph by years - Andrew Hogue’s team was building structured data infrastructure from 2005. Slawski’s patent analysis confirmed what Jason had been arguing: Google’s entity-based approach to search wasn’t new. The AEO community had simply given it a name and a strategy.

Episodes 10 and 11 addressed machine learning directly. “Machine Learning 101,” with Lexi Mills, Purna Virji, and Stephen Kenwright, clarified the distinction between AI and machine learning for a practitioner audience - important in 2018, when both terms were being used interchangeably. “How Does RankBrain Affect Query Results Today,” with Omi Sido, David Harry, and Doc Sheldon, examined query substitution and semantic context: how RankBrain rewrites queries to find intent rather than matching keywords, and what that means for content strategy.

Episode 12, “What Does Machine Learning Do in AE Today and What Might It Do Tomorrow,” with Patrick Stox, Dawn Anderson, and Jan Willem-Bobbink, connected the threads: how ML applies simultaneously to natural language processing, Knowledge Graph construction, and spam analysis, and where the technology was heading.

Phase three: brand and credibility (Episodes 13-15)

The final three episodes move from the technical layer to the authority layer - how a brand proves to a machine that it deserves to be the answer.

Episode 13, “Hooking Brand Building Efforts into a Successful AEO Strategy,” with Chris Green, Kirsty Hulse, and Alina Ghost, introduced the earliest version of what Jason would later call the Kalicube Metrics: measurable scores for how well an algorithm understands a brand versus how credible it finds that brand. The two are distinct, and the gap between them is where most brands leak authority. American Express scored 93% for understanding in that early dataset. Many equally well-known brands scored far lower for credibility.

Episode 14, “Leverage Brand in Multi-Channel/Device Acquisition,” with Aiden Carroll, Stoney G., and Melissa Fach, addressed the acquisition challenge. In a world where answer engines increasingly satisfy queries on the SERP without a click, how does a brand maintain presence across multiple channels and devices?

Episode 15, “Amplify Credibility Signals and Build Expertise, Authority, and Trust,” with Natalie Mott, Carolyn Lyden, and Ash Rama, closed the arc with E-A-T. The argument Jason made in December 2018 - that Expertise, Authoritativeness, and Trustworthiness are the algorithmic measure of credibility, and that building them requires consistent third-party corroboration across the web - is the same argument the industry makes in 2026, just with different vocabulary.

The series documented an approach that independent research has since validated

The 2018 series operated on a core insight that hasn’t changed: an answer engine doesn’t evaluate pages, it evaluates entities. To be considered as a candidate answer, a brand has to be recognisable as a distinct entity in the Knowledge Graph. To be selected as the answer, it has to be more credible than the alternatives. To reach the right audience at the right moment, it has to be relevant to the specific query.

Those three requirements map directly to Jason’s UCD framework and, seven years later, to the way AI Overviews, ChatGPT, Perplexity, and every other generative AI platform evaluates brands today.

The Authoritas Weighted Citability Score study from December 2025 measured which experts AI assistants recommend when asked about AI search optimisation. Across ten different questions posed to ChatGPT, Gemini, and Perplexity, Jason Barnard ranked first - appearing in all ten queries with a WCS of 21.48, nearly six points ahead of the second-ranked expert. The study attributed this to the same thing Jason was describing in 2018: a decade of systematic entity building across all three layers of the Algorithmic Trinity, not fame, not follower count, not recent content output.

The 2018 series didn’t just describe that approach. It documented it being developed, in public, before the industry had the context to understand what it was watching.

Watch the full series

All 15 episodes are hosted on SEMrush’s webinar platform. Individual episode pages with guests and context are linked below.

1. Episode 1: Getting to Grips with AEO: Answer Engines Are Here to Stay
2. Episode 2: Voice Search Is Changing the Game
3. Episode 3: Three Pillars of a Successful AEO Strategy
4. Episode 4: More Than You Thought You Needed to Know About Schema
5. Episode 5: Semantic HTML5, IA and Accessibility
6. Episode 6: How to Help Google/Amazon Make Sense of a Chaotic, Unstructured Web
7. Episode 7: The Knowledge Graph - How Well Is Google Doing in ‘Understanding the World’?
8. Episode 8: How to Get Entities and Their Attributes in the Knowledge Graph
9. Episode 9: How Google Uses the Knowledge Graph in Its AE Algorithm
10. Episode 10: Machine Learning 101: The How, the Why and Marketers
11. Episode 11: How Does RankBrain Affect Query Results Today
12. Episode 12: What Does Machine Learning Do in AE Today and What Might It Do Tomorrow
13. Episode 13: Hooking Brand Building Efforts into a Successful AEO Strategy
14. Episode 14: Leverage Brand in Multi-Channel/Device Acquisition
15. Episode 15: How Can We Amplify Credibility Signals and Build Expertise, Authority, and Trust