Answer Engine Optimization Glossary (2026)

The discipline of optimizing content to be cited by answer engines (Perplexity, ChatGPT, Google AI Overviews, Microsoft Copilot, Claude, Gemini). SEO ranks links; AEO earns citations inside the AI's answer.

Related: GEO, ASEO, Answer engine

A category of AEO signals measuring whether AI crawlers can fetch, parse, and understand your site. Includes llms.txt, llms-full.txt, llm.json, structured data, and clean static HTML.

Examples: GPTBot (OpenAI), ClaudeBot (Anthropic), PerplexityBot, Google-Extended, Bingbot, Applebot-Extended, Meta-ExternalAgent, Bytespider. Some fetch for training data, some for real-time queries.

Related: GPTBot, ClaudeBot, PerplexityBot

Composite score across 25+ signals in seven categories: AI Crawl Access, Content Quality, Product Clarity, Structured Data, Agent Readiness, Trust Signals, and EEAT. Score >85 = consistently cited; <50 = invisible to AI.

Examples: Perplexity, ChatGPT (with web search), Google AI Overviews, Google AI Mode, Microsoft Copilot, Google Gemini. Distinct from chatbots that don't cite sources.

Synonym for AEO. Less common in 2026 — most practitioners use AEO.

Headline metric for enterprise AI visibility programs. Tracked over time vs competitors. Above 60% = category leadership.

Related: Share of voice, Visibility score

Fires when a ChatGPT user asks a question that requires fetching live web content. Different from GPTBot (training) and OAI-SearchBot (ChatGPT Search index). Block this and you lose live citation visibility.

Related: GPTBot, OAI-SearchBot

When a third-party site (Reddit thread, blog, G2 listing) is cited by ChatGPT for a category prompt, and your brand isn't mentioned in that same answer, that's a citation gap. Most actionable signal in AEO — pitch the source or build content that displaces it.

AI cites amplitude.com/compare in 5 product-recommendation answers. You're not in that page → citation gap → pitch them or write a competing comparison.

Aggregating citation gaps across many prompts to identify the highest-priority sources to influence. Unique feature in some AEO platforms.

Crawls public content for Claude's training corpus. Anthropic also runs Claude-Web for real-time fetches.

Related: AI crawler, Claude-Web

Out of ~7 major AI engines, how many cite you for category prompts? Important secondary metric — high coverage means resilient visibility (no single engine dominates your traffic).

Google's quality framework, but equally important for AI visibility. AI engines weigh trust signals — testimonials, team pages, press mentions, customer counts — when deciding what to recommend.

Related: Trust signals

JSON-LD schema type that tells AI engines your page contains question-and-answer pairs. The single highest-ROI structured data type for AEO — Q&A maps 1:1 to how AI structures answers.

{ "@type": "FAQPage", "mainEntity": [{ "@type": "Question", ... }] }

Related: JSON-LD, Schema.org

Broader than AEO — includes optimization for any generative AI output, not just answer engines that cite sources. AEO is a subset of GEO. Most tactics overlap.

Related: AEO

Not a separate crawler — a token that lets sites disallow training-data use for Gemini and Vertex AI without affecting standard Google Search indexing.

Related: AI crawler

Fetches content to add to OpenAI's training corpus. Identifies as 'GPTBot/1.2'. Block via robots.txt to opt out of training data without affecting ChatGPT-User or OAI-SearchBot.

Related: AI crawler, ChatGPT-User

Format for embedding structured data in web pages, recommended by Schema.org. AI engines parse JSON-LD to understand page entities — products, organizations, articles, FAQs.

Related: Schema.org, FAQ schema

JSON variant of llms.txt — same intent (tell AI what your product is) but structured for programmatic consumption. Used by AI agents that need typed product data.

Related: llms.txt

Comprehensive (typically 500-2000 word) document with detailed feature descriptions, all pricing tiers, integrations, technical details. AI fetches this for deep-research queries.

Related: llms.txt

Plain-text Markdown file at the root of your domain. Lists product name, one-line description, key features, pricing, target audience. Sites with valid llms.txt get cited 3× more often.

# YourProduct
> One-sentence description
## Key Features
- ...

Related: llms-full.txt, llm.json

Per-prompt visibility metric. If 10 prompts are tracked and you're cited in 6, mention rate = 60%. Headline metric in most AEO dashboards.

Related: Visibility score

Builds the dedicated index ChatGPT Search uses. Distinct from GPTBot (training) and ChatGPT-User (real-time).

Fetches public content when a user asks Perplexity a question. Identifies as 'PerplexityBot/1.0'. Blocking it removes you from Perplexity entirely.

Related: AI crawler

Tier 1 = primary recommendation, Tier 2 = secondary mention, Tier 3 = listed alongside others, Tier 4 = absent. Higher tiers drive more click-through.

Related: Citation position

Examples: ChatGPT-User, Claude-Web, PerplexityBot, Meta-ExternalFetcher. Doesn't add to training data — only fetches in response to a specific user question. Block these and AI can't cite your live content.

Related: Training crawler

Cooperative project from Google, Microsoft, Yahoo, Yandex defining types like SoftwareApplication, Organization, FAQPage, Product. AI engines use Schema.org markup as a primary signal.

Even when cited, AI's tone matters. 'Best-in-class' (positive) drives more clicks than 'limited features' (negative). Track sentiment per provider over time.

When AI answers category-level prompts, what % of brand mentions are yours vs competitors? Tracked as a trend line — moving share of voice up = winning the category.

Related: Brand citation share

Specific prompt your tool runs on a schedule (e.g., daily/weekly) across multiple AI engines to track whether your brand is cited. Examples: 'best CRM for startups,' 'Stripe alternatives,' 'X vs Y'.

Examples: GPTBot, ClaudeBot, Google-Extended, Applebot-Extended, Meta-ExternalAgent, Bytespider. Block these to opt out of training without losing live citation visibility (allow real-time crawlers).

Related: Real-time crawler

Testimonials with named customers, user counts ('trusted by 5,000+ teams'), customer logos, press mentions, security/privacy pages, About pages with team bios, review platform listings (G2, Trustpilot).

Common formula: presence rate × 80% + cross-model consistency × 20%. Above 50% = healthy; below 25% = invisible. Different from AIES (which measures readiness) — visibility score measures actual outcomes.

Related: Mention rate, AIES