Skip to content
Anselis

Reference

Glossary

Short, plain-language definitions of the terms we use across this site — agent-ready, MCP, AEO, and the rest.

Agent

Software that uses an LLM to plan and execute multi-step tasks autonomously, calling tools and APIs to reach a goal.

Agentic economy

The emerging market in which AI agents — not only humans — discover, consume, and pay for digital services.

Agent-ready

A product is agent-ready when AI agents can discover it, understand what it does, and call its operations safely on a user's behalf. In practice: a machine-readable description of each capability (parameters, auth, error shapes, side-effects) plus validation, observability, and operation suitable for agent traffic.

AI Readiness

The set of capabilities (discoverability, semantic exposure, machine interfaces, governance) that make a product consumable by AI agents. Roughly synonymous with "agent-readiness", framed as a maturity property.

Protocol-agnostic

Property of a platform that does not depend on any specific agent protocol. MCP today, A2A or other future standards plug in over the same capability model and operation layer without rewriting the product.

AEO

Answer Engine Optimization — practices that make a website's content preferred and citable by LLM-based answer engines (ChatGPT, Claude, Perplexity, Gemini, Google AI Overviews).

GEO

Generative Engine Optimization — synonym of AEO, more commonly used in marketing circles.

LLM

Large Language Model — the model class powering current AI assistants and agents (e.g. Claude, GPT, Gemini).

MCP

Model Context Protocol — open protocol that standardizes how AI agents discover and use external tools, resources, and prompts.

MCP Server

A server that exposes a set of tools/resources to agents over MCP.

WebMCP

Extending the MCP idea to the open web: making a website expose its capabilities to agents through deterministic, machine-readable interfaces.

MCP Apps

Extension of MCP that lets a server return interactive UI components (widgets) the agent renders inline in the conversation — a calendar, a map, an editable quote — wired to the same tools/operations, so a capability can surface as a visual interface, not only text.

Tool (MCP)

A function an agent can call (e.g. `search_menu`, `create_booking`).

Resource (MCP)

Read-only data an agent can fetch (e.g. policies, catalogues).