DEFINITION REFERENCE MAY 2026·16 MIN READ

The Ecommerce AI Search Glossary: 50+ Terms Defined.

The complete vocabulary of AI search optimization for ecommerce brand owners. AISO, GEO, AEO, LLM, RAG, llms.txt, COSMO, Rufus, schema, E-E-A-T, Share of Model, answer capsules, every AI crawler, every measurement metric — all in one reference page.

50+Terms defined across 12 thematic categories
12Topic categories from foundational to technical
9AI crawlers covered in detail
100%Ecommerce-specific context for every term

AI search optimization picked up a vocabulary explosion between 2023 and 2026. New terms emerged faster than most brand owners could absorb them. This glossary is the complete vocabulary set, organized into 12 thematic categories, with definitions written specifically for ecommerce operators rather than academic researchers.

Each term is defined as a self-contained answer capsule (40-80 words) that explains both what the term means and how it applies to ecommerce in practice. Use this page as a reference while reading our other AI search guides — or send the link to a team member who needs to come up to speed quickly. For the broader strategic framework, see our What Is AI Search Optimization? page. For the academic-rooted GEO definition, What Is Generative Engine Optimization?.

01

Foundational AI Search Terms

The umbrella concepts that underpin everything else.

AISO — AI Search Optimization

The operator-friendly umbrella term covering all forms of AI search optimization across ChatGPT, Claude, Gemini, Perplexity, Google AI Overviews, and Amazon Rufus. AISO is the broadest of the related terms (GEO, AEO) and the one most marketing teams use day-to-day. See What Is AI Search Optimization?.

GEO — Generative Engine Optimization

The academic-origin term introduced in a November 2023 Princeton/Georgia Tech research paper. GEO refers specifically to optimization for generative AI engines (ChatGPT, Perplexity, Claude) and is essentially synonymous with AISO. Used most often in research and technical contexts. See What Is GEO?.

AEO — Answer Engine Optimization

The older term that predates ChatGPT, originally focused on featured snippets, voice assistants (Alexa, Google Assistant), and direct-answer extraction. AEO now refers to the subset of AISO focused on direct-answer formats. Increasingly subsumed by AISO/GEO terminology in 2026.

SEO — Search Engine Optimization

Traditional optimization for blue-link search engine results (Google, Bing). SEO competes for ranking in a list. AISO competes for citation in a generated answer. The two disciplines overlap heavily at the technical layer but play different content and authority games. Most ecommerce brands need both in 2026.

LLMO — Large Language Model Optimization

An emerging term referring to the optimization work that targets LLMs specifically (the underlying models behind AI search). Often used interchangeably with AISO and GEO. Less mainstream as a term but appears in technical and research contexts.

02

Generative AI Engines

The platforms ecommerce brands compete in for AI citations.

ChatGPT

OpenAI's flagship conversational AI assistant. The largest standalone generative engine with 700M+ weekly users in 2026. Heavily cites listicles (43.8% of citations), Reddit threads, YouTube transcripts, and structured guides. Optimization mechanic: third-party brand mentions, llms.txt configuration, Wikipedia/Wikidata authority, listicle inclusion.

Claude

Anthropic's AI assistant, smaller share of consumer ecommerce queries but disproportionate weight in B2B research and developer tools. Rewards technically clean sites with strong schema, named-author bylines, and well-structured comparison content. Same foundational optimization stack as ChatGPT.

Google Gemini

Google's AI assistant, integrated across the Google ecosystem (Workspace, Android, Chrome). Cites Google's broader index plus YouTube transcripts heavily. Foundational optimization (schema, content structure, E-E-A-T) covers Gemini alongside other engines.

Perplexity

AI search engine focused on cited, source-backed answers. Smaller raw user base but a high-intent audience including researchers, comparison shoppers, and B2B buyers. Cites primary sources heavily and rewards content with named authors and verifiable credentials. The 9 GEO research strategies (Cite Sources, Statistics Addition, Quotation Addition) lift visibility most reliably here.

Google AI Overviews (AIOs)

The AI-generated answer summaries that appear at the top of Google search results. AIOs appear on 50%+ of US searches and 14% of pure shopping queries. AIO trigger rate is 83% on "best [product]" queries. Optimized via FAQPage schema, Article schema, Product schema, clean Merchant Center feeds, and answer capsules under question-format H2s. See Google AI Overviews for Ecommerce.

Google AI Mode

Google's dedicated conversational AI search interface, launched in 2024 and separate from AI Overviews. Cites different sources than AIOs (only 13.7% citation overlap per Ahrefs data). AI Mode favors deeper, more comprehensive content; AI Overviews reward extractable passages. Treat as separate channels.

03

Amazon AI Terms

The Amazon-specific AI systems that drive in-platform shopping.

Amazon Rufus

Amazon's conversational shopping assistant launched in 2024 and now embedded across Amazon's mobile app and desktop experience. Rufus reads listing titles, bullets, A+ content, and descriptions to answer shopper questions and make recommendations. Optimizing for Rufus means clear noun-phrase optimization, complete A+ modules, robust review base, and structured product attribute coverage. See Amazon Listing Optimization.

COSMO

Amazon's commonsense knowledge graph and reasoning system, deployed across Amazon's search and recommendation infrastructure. Complements the older A9/A10 algorithm by adding contextual understanding. COSMO enables Rufus to match shopper intent to products even when keyword matches are imperfect. Listings optimized for COSMO use clear noun phrases, complete attributes, and specific use-case language.

A9 Algorithm

Amazon's original organic search ranking algorithm. Weights conversion rate, click-through rate, sales velocity, keyword match, and listing completeness to determine search ranking. Still active alongside COSMO in 2026 but less dominant than it was in 2018-2022.

A10 Algorithm

A semi-formal name for the evolved version of Amazon's search algorithm that includes more behavioral signals (off-Amazon traffic value, customer review quality, brand authority). Often used interchangeably with A9 in practice. The exact internal name Amazon uses is not publicly confirmed.

Noun-Phrase Optimization

The practice of using clear noun phrases ("insulated stainless steel water bottle") rather than disconnected keywords ("water, bottle, insulated") in Amazon listings. Critical for Rufus and COSMO optimization because the AI systems parse noun phrases as complete concepts. Listings written in clear noun-phrase prose perform measurably better than keyword-stuffed listings in 2026.

04

Citation & Visibility Metrics

The new metrics that measure AISO success.

Citation Rate

The percentage of target category queries where your brand is named in the AI-generated response. Tracked by running 30-50 representative queries weekly across each engine and counting where your brand appears. The most direct measure of whether your AISO investment is working.

Share of Model (SOM)

Your citation rate divided by the total citations across you plus top 3-5 competitors in the same query set. The cleanest competitive AISO benchmark, analogous to Share of Voice in traditional advertising. Tracked monthly to measure relative AI visibility against the competitive set.

Citation Position

When your brand is cited, where in the AI response does it appear? First-mention citations earn the most click-through; later-position citations still build brand awareness but drive less direct traffic. Track citation position alongside citation rate for full picture.

AI Referral Traffic

Sessions arriving at your site from AI engines (ChatGPT, Perplexity, Google AI Overviews). Tracked via custom GA4 channel groupings. AI referral traffic typically converts at 3-5x the rate of traditional organic traffic, making it disproportionately valuable per visit.

AIO Impressions

Google Search Console added AI Overview impressions as a separate data point in 2025. Found in the Performance report. Measures how often your domain appears as a citation source within AIOs. Best paired with click data to understand which AIO citations drive traffic vs purely brand awareness.

Branded Search Lift

An indirect AISO success metric. As AI citations build awareness, branded search volume on Google rises 2-3 months later. One of the cleanest leading indicators of AISO compounding into broader brand awareness.

Evolve Media Agency Service

AI Search Optimization for Ecommerce

We deploy the full AISO stack — schema, llms.txt, answer capsules, Wikidata, third-party mentions — for $1M-$10M ecommerce brands.

Book a Free Call →
Free Resource

The Ecom Profit Box

11 step-by-step PDF guides covering launches, content, split testing, email flows, and AI search foundations.

Grab It Free →
05

Schema & Structured Data

The structured markup AI engines parse for context.

Schema.org

The collaborative vocabulary maintained by Google, Microsoft, Yahoo, and Yandex for structured data on the web. Schema.org defines the types (Article, Product, FAQPage, Person, etc.) and properties (name, description, dateModified) that make web content machine-readable. AI engines parse schema to understand what content is about and how authoritative it is.

JSON-LD

The JSON-based format for embedding schema markup in web pages. Placed in a script tag in the HTML head. The recommended format for schema implementation in 2026 because it is easier to maintain than embedded microdata or RDFa.

Article Schema

Structured data for editorial content. Includes headline, datePublished, dateModified, author (Person), publisher (Organization), and mainEntityOfPage. Required schema for any cornerstone content page. Without Article schema, AI engines cannot confidently attribute content to a verifiable author or organization.

FAQPage Schema

Structured data for FAQ sections, marking each question and answer pair. Pages with FAQPage schema are cited 2.8x more often in AI responses than equivalent unmarked content — the highest citation multiplier of any single schema type. Every cornerstone page should include 6-12 FAQ entries with FAQPage schema.

Product Schema

Structured data for product pages. Required fields for ecommerce: name, description, image, brand, offers (price, currency, availability), aggregateRating, and review. Feeds into Google's Shopping Graph for AI Overview product recommendations.

BreadcrumbList Schema

Structured data marking the navigation hierarchy of a page (Home → Category → Subcategory → Page). Helps AI engines understand site structure and content relationships. Foundational schema that should appear on every page.

Organization Schema

Structured data establishing your brand as a verifiable entity. Includes name, logo, sameAs (social profile chain), founder, founding date, address, and contact information. Should appear on the homepage at minimum. The foundation for entity-based AI visibility.

Person Schema

Structured data for named individuals (typically authors). Includes name, sameAs chain (LinkedIn, Twitter, author page), credentials, and affiliation. Essential for E-E-A-T signal strength. AI engines weight named-author content significantly higher than anonymous content.

HowTo Schema

Structured data for instructional content with numbered steps. Includes step descriptions, required tools, estimated time, and total time. Particularly powerful for "how to use [product]" and "how to set up [category]" content because it explicitly signals to AI that the content contains a structured answer.

06

Technical AISO Foundation

The technical infrastructure that makes AI citation possible.

llms.txt

An emerging standard text file placed at the root of a domain (similar to robots.txt) that tells AI systems what your site is about and which pages are most important. Lists key pages, brand description, and content priority. Not yet officially mandated by AI engines but widely supported. We recommend deploying both llms.txt and llms-full.txt as foundational AISO work. See llms.txt Guide for Ecommerce.

llms-full.txt

The longer companion to llms.txt that includes more detailed brand context, full descriptions of key pages, and complete entity information. Where llms.txt is a brief summary, llms-full.txt is comprehensive. Both files placed at the root domain.

robots.txt

The text file at the root of a domain that tells crawlers (search engines and AI bots) which parts of the site they can or cannot access. Critical for AISO because blocking AI crawlers in robots.txt eliminates your visibility on those platforms. In 2026, ecommerce brands should explicitly allow GPTBot, ClaudeBot, PerplexityBot, Google-Extended, OAI-SearchBot, ChatGPT-User, and Amazonbot.

Server-Side Rendering (SSR)

Web architecture where page content is rendered on the server before being sent to the browser. The opposite of client-side rendering (where JavaScript runs in the browser to assemble the page). AI crawlers struggle with JavaScript-heavy sites that require browser execution to display content. SSR ensures AI engines can read your content reliably.

Core Web Vitals

Google's set of page experience metrics: Largest Contentful Paint (LCP), Cumulative Layout Shift (CLS), and Interaction to Next Paint (INP). Now extended with First Contentful Paint (FCP) for AI search optimization. Pages with FCP under 0.4 seconds are cited 3x more often than slower pages in AIO research. Foundational technical signal for AISO.

Structured Data Testing

Verifying that schema markup is correctly implemented and parseable. Tools include Schema.org Validator, Google's Rich Results Test, and the Schema.org JSON-LD Playground. Schema with errors fails to provide visibility lift even when present in the HTML.

07

Content Strategy Terms

The content patterns that drive AI citation.

Answer Capsule

A 40-60 word self-contained passage placed directly under an H2 or H3 heading that completely answers the question posed by the heading. The single most important content format for AI citation because AI engines extract passage-sized chunks from web pages rather than entire pages. Answer capsules are written so they make sense in isolation, include specific entities and numbers, and front-load the answer.

Passage Extraction

How AI engines select specific 134-167 word self-contained passages from web pages to include in generated answers. AI engines do not cite entire pages — they cite quotable chunks. Content scoring high on semantic completeness is 4.2x more likely to be cited per Wellows research.

Semantic Completeness

The measure of how thoroughly a passage answers a query without requiring surrounding context. The #1 AI citation ranking factor (r=0.87 correlation in research). Content scoring 8.5/10 or higher on semantic completeness is 4.2x more likely to earn an AI citation. Tools like NeuronWriter, Surfer SEO, and Clearscope can score content against semantic completeness baselines.

Question-Format H2

An H2 heading written as a question matching conversational query patterns ("How does X work" instead of "X overview"). Question-format headings match the way users phrase queries to AI engines and trigger passage extraction more reliably than declarative headings.

Listicle

Content structured as a numbered or itemized list, typically "Best [X]" or "Top N [Y]" format. The single highest-AIO-trigger content type for commercial queries. AIO triggers on 83% of "best [product]" queries, and 43.8% of ChatGPT citations come from listicle pages per Ahrefs data.

Comparison Page (Versus Page)

Content explicitly comparing two or more products, services, or categories ("[Brand A] vs [Brand B]"). Comparison queries trigger AIOs at approximately 65% rate. Pages with 3+ comparison tables earn 25.7% more AIO citations per AirOps research.

Buying Guide

Long-form content (typically 2,000-4,000 words) helping shoppers research a category before purchase. Triggers AIOs on roughly 58% of "how to choose [X]" queries. Highest-leverage ecommerce content format when combined with full schema markup and answer capsules.

Definitional Page

Content answering "what is [X]" queries with comprehensive definitions, examples, and context. Triggers AIOs on roughly 54% of definitional queries. Examples include this glossary, our What Is AISO page, and our What Is an Amazon Agency page.

08

Authority & E-E-A-T

The credibility signals AI engines weight when deciding who to cite.

E-E-A-T

Stands for Experience, Expertise, Authoritativeness, and Trustworthiness. Google's framework for evaluating content quality, now actively used by AI search systems to decide which sources to cite. Strong E-E-A-T signals include named-author bylines with credentials, original first-hand content, citations of authoritative sources, third-party brand mentions, and verifiable contact information. See E-E-A-T for Ecommerce.

Wikidata

The free, collaborative knowledge graph maintained by the Wikimedia Foundation. Stores structured data about brands, people, products, and concepts. AI engines use Wikidata as a foundational source of entity recognition. Creating Wikidata entries for your brand and founder is one of the highest-leverage AISO authority moves available.

Wikipedia

The collaborative encyclopedia. Wikipedia is the #1 cited source in Google AI Mode (11.22% of mentions per Ahrefs) and a heavily weighted source across all AI engines. Notability requirements make Wikipedia inclusion difficult for most ecommerce brands, but founders and brands with sufficient third-party press coverage can earn entries.

sameAs Property

A schema.org property that links your Person or Organization entity to verified profiles on external platforms (LinkedIn, Twitter, Crunchbase, Wikidata, Wikipedia). The sameAs chain is what AI engines follow to verify your entity is real and consistent across the web. Critical for entity authority signals.

Entity Authority

The strength of recognition AI engines have for your brand or person as a verifiable entity. Built through Wikidata entries, consistent profile data across platforms, named-author bylines with Person schema, and third-party mentions that corroborate your existence. Brands with strong entity authority get cited by name; brands with weak entity authority get described generically.

Third-Party Brand Mentions

References to your brand in content you do not control — podcast episodes, listicles on independent sites, Reddit threads, news articles, HARO/Qwoted responses. AI engines weight third-party mentions heavily because they corroborate your brand exists outside your own marketing. See Brand Mention Strategy for AI Search.

09

AI Crawlers & Bots

The specific user-agents that crawl your site for AI training and search.

GPTBot

OpenAI's crawler used to gather training data for GPT models. Identified by the user-agent "GPTBot." Blocking GPTBot prevents your content from being included in OpenAI training datasets but does not block real-time ChatGPT browsing.

OAI-SearchBot

OpenAI's crawler for ChatGPT search functionality, separate from GPTBot. Used to populate ChatGPT's search index. Allowing OAI-SearchBot is necessary for visibility in ChatGPT's search-augmented responses.

ChatGPT-User

The user-agent ChatGPT uses when fetching pages in real-time during a user conversation (e.g., when a user asks ChatGPT to "look at this page"). Allowing ChatGPT-User enables your pages to be retrievable when users specifically reference them.

ClaudeBot

Anthropic's crawler used to gather training data and populate Claude's knowledge. Blocking ClaudeBot eliminates your visibility on Claude. Allow in robots.txt for ecommerce AISO.

PerplexityBot

Perplexity's crawler. Critical to allow because Perplexity's high-intent comparison-shopping audience converts well for ecommerce. Blocking PerplexityBot eliminates one of the highest-quality AI traffic sources.

Google-Extended

Google's separate user-agent for AI training data collection (Bard/Gemini). Distinct from Googlebot (which crawls for traditional Google Search). Allowing Google-Extended is necessary for inclusion in Gemini's knowledge base.

AppleBot-Extended

Apple's user-agent for Apple Intelligence AI training. Newer crawler launched alongside Apple Intelligence in 2024-2025. Allow in robots.txt to maintain visibility as Apple's AI features expand.

Meta-ExternalAgent

Meta's user-agent for Meta AI training. Allow for visibility in Meta AI features across Instagram, Facebook, and WhatsApp.

Amazonbot

Amazon's crawler used to gather information for Alexa, Rufus, and other Amazon AI features. Critical for Amazon sellers to allow because Amazonbot feeds Rufus's knowledge of off-Amazon content related to your products.

The fastest AISO upgrade most ecommerce brands can make takes ten minutes: open robots.txt, allow every AI crawler, save. The compounding visibility lift starts the same day.
— Ian Smith, Founder, Evolve Media Agency
10

Google Shopping & Merchant

The Google-specific commerce infrastructure for AI Overviews.

Google Shopping Graph

Google's structured database of products and merchants that powers product recommendations within AI Overviews for commercial queries. A properly configured Merchant Center feed is the entry ticket to Shopping Graph inclusion. Missing or erroneous feed data eliminates AIO product visibility regardless of SEO strength.

Google Merchant Center

The platform where ecommerce brands submit product feed data to Google for inclusion in Google Shopping, AIO product recommendations, and Performance Max campaigns. Required fields include brand, material, color, audience, aggregateRating, and complete product specifications. Use Google Merchant Center Next (the 2024+ version) for the current feature set.

Product Feed

The structured file (typically XML or Google Sheets) that ecommerce platforms use to send product data to Merchant Center, Meta Catalog, TikTok Shop, and other commerce surfaces. Variant consistency across page HTML, JSON-LD schema, and product feed is critical — discrepancies cause AI engines to lose trust in the data.

AIO Triggering

The condition under which a search query produces an AI Overview vs traditional blue-link results. Triggering varies by query type: 83% on "best [product]" queries, ~65% on comparisons, ~58% on buying guides, ~5% on pure transactional queries. Mapping your target keywords against AIO trigger rates helps prioritize content investment.

Variant Consistency

The requirement that product variant data (size, color, material) match exactly across page HTML, JSON-LD schema, and Merchant Center feed. Discrepancies cause AI engines to lose trust in the data and skip citation. Audit variant consistency before launching new product pages.

11

Measurement & Analytics

The tracking infrastructure for AISO success.

Manual Prompt Testing

The most accurate baseline citation tracking method. Run 30-50 representative category queries weekly across each AI engine (ChatGPT, Perplexity, Gemini, AI Overviews) and document where your brand appears. Cheapest and most reliable measurement approach. Tool-based tracking automates this but with some accuracy tradeoffs.

30-Query Prompt Test

The standard measurement approach: a curated set of 30 representative category queries run consistently across engines on a weekly or monthly basis. Documents Citation Rate, Citation Position, and Share of Model. Forms the baseline against which AISO improvements are measured.

GA4 Custom Channel Grouping

The setup in Google Analytics 4 that captures AI referral traffic as a distinct channel. Required because AI engines often appear under "Direct" or "Other" without custom configuration. Set up referral filters for chatgpt.com, perplexity.ai, and AI-related Google referrer parameters.

Search Console AI Overview Filter

The 2025 addition to Google Search Console that lets you filter Performance data by AI Overview impressions. Available in the search appearance filter. Critical for understanding which keywords drive AIO impressions vs traditional organic.

Conversion Rate Premium

The measurable difference in conversion rate between AI referral traffic and traditional organic traffic. Industry data shows AI Overview traffic converts at 14.2% vs 2.8% for traditional organic — a 5x premium. Justifies higher cost-per-visit on AISO investment because each visit is worth more.

12

Tools & Platforms

The software stack ecommerce brands use for AISO work.

Otterly.ai

Automated citation tracking platform monitoring brand appearances across ChatGPT, Perplexity, and Gemini. Tracks Citation Rate and Share of Model across hundreds of keywords. Mid-tier pricing suitable for $1M-$10M brands.

Profound

Enterprise AI visibility platform with brand monitoring, competitor benchmarking, and citation analytics. Higher pricing tier suited to $10M+ brands or agencies managing multiple clients.

Semrush AI Visibility Toolkit

AI search optimization features integrated into the Semrush platform. Suitable for brands already using Semrush for traditional SEO who want AISO tracking in the same workflow.

NeuronWriter

Content optimization tool that scores draft content against semantic completeness and competing pages. Strong for ecommerce content because it analyzes Google AIO citation patterns directly.

Surfer SEO

Content optimization platform with AI search optimization features. Scores content against ranking factors and provides keyword and entity recommendations.

Clearscope

Content grading platform focused on semantic completeness and topical authority. Premium pricing tier suited to publishers and content teams.

Schema.org Validator

The official schema implementation testing tool from Schema.org. Verifies that JSON-LD markup is correctly structured and parseable. Free.

Google Rich Results Test

Google's official tool for testing structured data eligibility for enhanced search features. Free. Use alongside Schema.org Validator for full schema validation.

WPCode

WordPress plugin for deploying schema snippets, header scripts, and custom code without theme edits. Standard tool for ecommerce AISO implementation on WordPress. Used heavily for Yoast schema overrides and custom schema deployment.

Yoast SEO

WordPress SEO plugin that handles default schema generation, meta tags, sitemaps, and OG cards. Provides foundational SEO/AISO infrastructure on WordPress sites. Often used alongside WPCode for custom schema additions.

Common Questions

AI Search Vocabulary
FAQ

What is the difference between AISO, GEO, and AEO?

AISO (AI Search Optimization) is the operator-friendly umbrella term covering all AI search systems. GEO (Generative Engine Optimization) is the academic-origin term from a 2023 Princeton/Georgia Tech research paper, focused specifically on generative engines. AEO (Answer Engine Optimization) is the older term that predates ChatGPT, now used for the subset of AISO focused on direct-answer extraction. The three terms describe overlapping work but with different scopes — AISO is broadest, GEO is most academic, AEO is narrowest.

What is an LLM and how does it relate to AI search?

An LLM (Large Language Model) is a type of artificial intelligence trained on vast amounts of text to understand and generate human-like language. Examples include GPT-4, Claude, Gemini, and the underlying models powering ChatGPT, Google AI Overviews, and Perplexity. AI search systems use LLMs to process queries, retrieve relevant information, and synthesize answers. When ecommerce brands optimize for AI search, they are optimizing for which information LLMs cite when generating answers.

What is RAG (Retrieval-Augmented Generation)?

RAG is the architecture powering most modern AI search systems. When a user asks a question, RAG-based systems first retrieve relevant content from indexed sources (the open web, Wikipedia, YouTube transcripts, structured databases), then use a language model to synthesize a generated answer drawing from the retrieved content. RAG lets AI engines cite specific sources rather than just generating from training data alone, which is why ecommerce brands can earn citations through content optimization.

What is COSMO on Amazon?

COSMO is Amazon's commonsense knowledge graph and reasoning system, deployed across Amazon's search and recommendation infrastructure. It complements the older A9/A10 search algorithm by adding contextual understanding. COSMO is what enables Rufus to understand that a query like "something to keep my drink cold for hiking" relates to insulated water bottles even without explicit keyword overlap. Listings optimized for COSMO use clear noun phrases, complete attribute coverage, and specific use-case language.

What is llms.txt and is it required?

llms.txt is an emerging standard text file placed at the root of a domain (similar to robots.txt) that tells AI systems what your site is about and which pages are most important. It is not yet required — no AI engine officially mandates it. However, several major AI systems have signaled support for it, and brands deploying llms.txt early gain a small but consistent advantage. We recommend deploying llms.txt and llms-full.txt as part of foundational AISO work in 2026.

What is Share of Model?

Share of Model (SOM) is the percentage of category-relevant AI responses that name your brand. It is calculated by running 30-50 representative category queries across multiple AI engines, counting how often your brand appears in the responses, and dividing by the total. SOM is the cleanest competitive benchmark for AI search visibility — similar to how Share of Voice works in traditional advertising. Tracking SOM monthly across competitors reveals whether your AISO investment is translating into real category presence.

What is an answer capsule?

An answer capsule is a 40-60 word self-contained passage placed directly under an H2 or H3 heading that completely answers the question posed by the heading. It is the single most important content format for AI citation because AI engines extract passage-sized chunks from web pages rather than entire pages. Answer capsules are written so they make sense in isolation, include specific entities and numbers, and front-load the answer rather than building up to it through context.

Which AI crawlers should I allow in robots.txt?

The major AI crawlers ecommerce brands should explicitly allow as of 2026: GPTBot (OpenAI training), OAI-SearchBot (ChatGPT search), ChatGPT-User (in-session ChatGPT browsing), ClaudeBot (Anthropic), PerplexityBot (Perplexity), Google-Extended (Google AI training), AppleBot-Extended (Apple Intelligence), Meta-ExternalAgent (Meta AI), and Amazonbot (Amazon search). Blocking any of these eliminates your visibility on the corresponding platform. Most ecommerce brands should allow all AI crawlers — the visibility upside dramatically outweighs the training-data concerns.

What is E-E-A-T and why does it matter for AI search?

E-E-A-T stands for Experience, Expertise, Authoritativeness, and Trustworthiness. It is Google's framework for evaluating content quality and is now actively used by AI search systems to decide which sources to cite. Strong E-E-A-T signals include named-author bylines with credentials, original first-hand content (photography, video, case studies), citations of authoritative sources, third-party brand mentions, and verifiable contact information. Pages with strong E-E-A-T signals are cited 2.3x more often than pages with weak signals.

What is passage extraction in AI search?

Passage extraction is how AI search engines select specific 134-167 word self-contained passages from web pages to include in generated answers. AI engines do not cite entire pages — they cite quotable chunks. Content scoring high on semantic completeness is 4.2x more likely to be cited per Wellows research. For ecommerce brands, this means writing answer capsules under each H2 heading and ensuring those passages can stand alone as complete answers.

Ian Smith, Founder of Evolve Media Agency
Ian Smith
Founder, Evolve Media Agency · Ecommerce & AI Search Specialist

Ian co-founded Evolve Media Agency with his wife Megan in 2017. Based in Colorado, Evolve Media is a small boutique team serving $1M-$10M ecommerce brands across photo, video, listing optimization, AI search, and growth strategy. Ian has spent the last two years deep-diving into GEO and AI search across ChatGPT, Claude, Gemini, and Perplexity. Read Ian's full bio →

Work With Ian

Apply The Vocabulary, Win The Citations

Ready to Implement AI Search for Real?

Knowing the terms is the first step. Implementing the full AISO stack is the harder one. Evolve Media Agency builds the technical foundation, content structure, schema, and authority layer for $1M-$10M ecommerce brands. Small Colorado team, low overhead, real client relationships, fast turnaround.