Key concepts & glossary
GEO, AEO, agent.json, llms.txt, MCP, trust.json, and the discoverability score explained.
Definitions for the terms that recur across these docs. Each links to the deeper page where the concept is used.
At a glance
- A glossary of the terms used across these docs, each linking to the page where it's applied.
- Mission pillars: Discover, Evaluate, and Trust — the three stages of the buyer journey.
- Optimization concepts: GEO, AEO, the capsule format (~40–60 word answers), and entity density.
- Machine-readable outputs:
agent.json,llms.txt, MCP, andtrust.json— served per company at/company/{slug}/.... - Knowledge structure: the five GEO sections (Discover, Educate, Evaluate, Convert, Enable) and the discoverability score.
Mission pillars
Discover
The first stage of the buyer journey: finding the AI companies and solutions that fit a problem, industry, or use case. On the company side, discovery work means a clear profile and well-listed solutions.
Evaluate
The comparison stage: judging options on features, pricing, case studies, and how AI assistants describe each brand. The AI Score grader and competitor comparisons live here.
Trust
The commitment stage: confirming a vendor is credible. Covered by the trust profile — governance, models in use, and certifications.
Optimization concepts
GEO (Generative Engine Optimization)
Structuring content so AI engines (ChatGPT, Perplexity, Gemini, and others) select, ground on, and cite it accurately when answering questions. GEO is an extension of SEO for the era where AI assistants answer directly instead of listing links.
AEO (Answer Engine Optimization)
A closely related discipline focused specifically on getting cited inside answer engines. In practice the strongest lever is a direct, self-contained answer placed right after a question or heading. Welcome.AI's company knowledge is structured as question-and-answer pairs for exactly this reason.
Capsule format
A short, self-contained answer of roughly 40–60 words placed immediately after a question or heading. Capsules are the format AI engines extract and cite most often, because they can be lifted whole without surrounding context. Knowledge answers are written to this length.
Entity density
Naming specific brands, tools, people, and products explicitly instead of speaking in vague generalities. High entity density reduces ambiguity for AI engines and makes your content easier to attribute and cite.
Machine-readable outputs
agent.json
A structured JSON profile of your company served at /company/{slug}/agent.json. It is designed for AI agents, retrieval systems, and competitive analysis tools to consume directly. The shape follows the AI Entity Spec. Response detail scales with your plan — name and tagline on free, the full profile (solutions, all five knowledge sections, competitors, and more) on Pro. See AI presence.
llms.txt
A plain-text, LLM-readable summary of your company served at /company/{slug}/llms.txt. It is the human- and model-readable companion to agent.json, following the emerging llms.txt convention. See AI presence.
MCP (Model Context Protocol)
An open standard for giving AI assistants structured access to external context and tools. Welcome.AI's machine-readable endpoints align with this direction so agents can pull your company data through a standardized interface. Covered alongside the other endpoints in AI presence.
trust.json
A self-reported, agent-readable trust record covering AI governance, the models a company uses, and its certifications. It lets an AI agent verify credibility programmatically rather than parsing marketing copy. See trust profile.
Knowledge structure
The five GEO knowledge sections
Company knowledge is organized as Q&A across five sections that follow the buyer journey:
| Section | Question it answers |
|---|---|
| Discover | What is this company and what does it do? |
| Educate | How does it work and what should I understand first? |
| Evaluate | How does it compare, and is it right for me? |
| Convert | What does it cost and how do I get started? |
| Enable | How do I adopt, integrate, and succeed with it? |
Each section holds capsule-length answers. See company knowledge.
Discoverability score
A composite readiness score across the discover, evaluate, and trust pillars. It tells you how complete and machine-readable your presence is, and what to fix next. See discoverability score.
FAQ
What's the difference between GEO and AEO?
GEO (Generative Engine Optimization) is the broad practice of structuring content so AI engines select, ground on, and cite it. AEO (Answer Engine Optimization) is the closely related, narrower focus on getting cited inside answer engines, where the strongest lever is a direct, self-contained answer placed right after a question.
What is the capsule format?
A short, self-contained answer of roughly 40–60 words placed immediately after a question or heading. AI engines extract and cite capsules most often because they can be lifted whole without surrounding context.
What are the five GEO knowledge sections?
Discover, Educate, Evaluate, Convert, and Enable — a Q&A structure that follows the buyer journey, answering what a company does, how it works, how it compares, what it costs, and how to adopt it.
What's the difference between agent.json and llms.txt?
agent.json is a structured JSON profile of a company at /company/{slug}/agent.json for agents and retrieval systems, while llms.txt is its plain-text, LLM-readable companion at /company/{slug}/llms.txt. See AI presence.