Canonical Terminology Index

Agentic Commerce Citation Glossary.

Sixteen concise, machine-readable definitions covering Universal Commerce Protocol (UCP), Agentic Commerce Protocol (ACP), Agent Payments Protocol (AP2), generative search optimization, and deterministic revenue infrastructure.

Universal Commerce Protocol: An open discovery standard served at /.well-known/ucp that publishes a merchant identity, structured product or service catalog, SLA commitments, and supported agent transport bindings (REST, MCP, A2A) in machine-readable JSON.
Agentic Commerce Protocol: A stateful machine-to-machine transaction specification (/api/acp/negotiate and /api/acp/checkout) enabling autonomous buyer agents to request deterministic quotes and execute idempotent checkouts.
Agent Payments Protocol: A cryptographic authorization layer (/.well-known/ap2/mandates.json) using ECDSA P-256 JSON Web Signatures (ES256) to prove principal consent, enforce spending caps, and provide non-repudiation for agent-executed payments.
Answer Engine Optimization: The discipline of structuring web pages with concise 40-to-60-word direct answers, semantic HTML headings, SpeakableSpecification, and FAQPage schemas so AI assistants extract and cite exact passages.
Generative Engine Optimization: Engineering verifiable fact density, HTML comparison tables, transparent pricing specifications, and corroborated entity graphs so generative models include a brand when synthesizing multi-criteria vendor recommendations.
Agentic Commerce Readiness Audit: A nine-pillar diagnostic scoring any domain across UCP, ACP, AP2, JSON-LD @graph completeness, server-rendered HTML latency (<300ms p95), AI bot access, llms.txt, entity provenance, and off-site citation corroboration.
Intent Mandate: An AP2 signed credential issued by a human or corporate principal pre-authorizing an autonomous agent to negotiate and purchase within defined merchant categories, spending ceilings, and expiration timestamps.
Cart Mandate: An AP2 cryptographic signature over exact finalized SKUs, quantities, currency, and settlement totals immediately prior to autonomous checkout execution.
Deterministic Quoting Engine: An architecture that restricts language models to structured scope intake while computing all prices, discounts, and margin floors inside deterministic TypeScript and PostgreSQL rules with zero arithmetic hallucination.
Sub-60-Second Speed-to-Lead: Engaging every inbound telephone or web form inquiry within 45 to 60 seconds via 24/7 autonomous voice and booking agents, capturing buyers at peak intent before lead decay sets in.
Waterfall Enrichment: Querying multiple firmographic, technographic, and verified-contact providers sequentially so downstream data sources are queried and billed only when higher-priority sources miss.
JSON-LD Entity Graph (@graph): A unified Schema.org linked-data payload connecting Organization, Person, WebSite, WebPage, Service, Article, and FAQPage nodes through persistent @id URIs.
llms.txt: A concise, root-level markdown index (/llms.txt and /llms-full.txt) designed specifically for low-token ingestion by large language models and autonomous research agents.
300ms AIO Discoverability Budget: The target 95th-percentile time-to-complete-HTML threshold ensuring real-time AI search crawlers receive full semantic copy and schemas without timing out or waiting on client JavaScript.
Model Context Protocol: A standardized tool and context interface allowing AI hosts and autonomous agents to invoke structured domain capabilities such as catalog search, readiness probes, and quote calculators.
Progressive WebGL Enhancement: Serving complete server-rendered HTML and ray-traced Cycles PNG posters immediately, then mounting interactive Three.js 3D scenes during browser idle periods when GPU and motion preferences permit.