Product truth · agent-readable definitions
The authoritative vocabulary for 3Dogs products, interfaces and roles. Pointer points. Authorized people and tools act. The Executive Committee proofs the work. 3Dogs packages the process.
The diagnostic and verification product. Pointer examines a supplied strategic question and AI answer, then points to evidence, conflicts, assumptions, gaps, risks and a suggested next human action. It does not execute the action. Verify My AI is a Pointer capability; A2A is Pointer's machine-readable interface.
The deeper DiscoveryâNexusâEvolution decision-intelligence system. It reframes the question, gathers current cited evidence, convenes independent AI models in structured debate, and produces one documented recommendation with confidence, risk and dissent preserved.
The authorized hands-on implementation role in the 3Dogs service process. It may inspect, change, test and document systems only within the authority granted by the customer or operator.
The supervisory and quality-assurance role. The committee challenges assumptions, reviews evidence, records dissent and proofs material work before approval or promotion.
The governed packaging of discovery, diagnosis, authorized implementation, independent review, evidence capture and continued improvement. Customer systems remain authoritative.
Agent2Agent Protocol: an open protocol that lets compatible agents discover one another, exchange tasks and receive structured results. A2A is an interface category, not a separate 3Dogs product. At 3Dogs, it is a machine-readable entry point to Pointer.
How well a public site can be discovered, understood, cited and safely used by AI agentsâincluding crawl policy, structured data, accessible content, Markdown mirrors, discovery files, API contracts and human-control boundaries.
The broad commercial ability for an AI assistant to find an organization, understand and cite its current public offerings and policies, and identify the safe actions available next. AI discovery is a capability category, not a separate 3Dogs product.
The technical process by which one agent finds another agent and understands its declared identity, endpoint, skills, capabilities and authentication requirements. A2A Agent Cards are one standardized mechanism.
The part of a buying journey in which an assistant interprets the customerâs intent and constraints, then finds and compares relevant current offers. It is more specific than the broader term AI discovery.
A buying or booking journey in which an AI assistant helps discover, compare and prepare a transaction. The seller and connected systems remain authoritative, and consequential authorization, terms and payment stay under human control unless a separately verified mandate explicitly allows otherwise.
The authorization and settlement portion of agentic commerce. It requires secure, auditable and bounded authority; discovery metadata or a caller-supplied consent flag is not payment authorization.
A legal term for identifying, collecting and producing electronically stored evidence. It is not the term for AI shopping, agent visibility or machine-readable business discovery.
Answer Engine Optimization: structuring accurate, direct and sourceable answers so answer engines can interpret and cite them.
Generative Engine Optimization: making authoritative content legible and useful to generative AI systems while preserving provenance and product truth.
Search Engine Optimization: improving the technical and content signals that help traditional search engines crawl, index and rank public pages.
Model Context Protocol: an open protocol for connecting AI applications to tools, resources and prompts through a defined server interface.
Agentic Commerce Protocol: an evolving machine-readable commerce specification for compatible applications and sellers. Its presence does not by itself prove conformance or authorize an agent to accept terms or make a payment for a human.
Universal Commerce Protocol: an evolving open commerce standard designed to connect consumer surfaces, businesses and payment providers across discovery, checkout and order-management capabilities. Publishing a profile does not by itself prove conformance.
An agent may explain options, prepare a request or obtain bounded authorization, but a person remains responsible for consequential decisions, applicable terms and payment unless a separately verified mandate explicitly allows otherwise.