AITWIRE Canonical Authority Specification

1. Overview

AI Systems increasingly answer questions about businesses — citing hours, prices, policies, and product details. When these answers are wrong, customers receive incorrect information. AITWIRE exists to solve this problem by giving businesses an Authority Graph — a structured, verifiable, machine-readable declaration of their canonical data.

The Authority Graph is published at a well-known endpoint (/.well-known/entity.json) and referenced from robots.txt, ai.txt, and llms.txt. AI crawlers and retrieval systems can discover it automatically and use it as the definitive source of truth.

2. The Authority Graph

An Authority Graph is a per-tenant document that declares:

• Surfaces — the digital properties a business controls (website, social profiles, marketplaces, listings)
• Proofs — cryptographic or DNS-based evidence that the business owns those surfaces
• Authority Rules — which surfaces are authoritative for which claim types, with conflict resolution policies
• Feeds — structured data endpoints for products, articles, locations, services, and policies

3. Surfaces

A surface is any digital property that publishes information about a business. Each surface has:

• surface_id — a unique identifier
• type — one of: website, social, marketplace, listing, review, directory
• uri — the canonical URL of the surface
• role — primary (authoritative source) or secondary (mirror/syndicated)
• status — active, revoked, or pending

4. Proofs & Verification

Proofs establish that a business controls a surface. AITWIRE supports four proof methods:

• DNS TXT — add a TXT record to the domain (_aitwire-verify.example.com)
• Meta tag — add a <meta name="aitwire-verify"> tag to the page
• JSON-LD — include an AITWIRE verification block in the page's structured data
• .well-known — place a verification file at /.well-known/aitwire-verify.json

Proofs have expiration dates and can be revoked. Expired proofs are automatically flagged for renewal.

5. Claim Types

Claim types classify the kinds of information that surfaces publish. The taxonomy includes:

operational.hours       — business hours
operational.location    — physical address
operational.contact     — phone, email, chat
policy.privacy          — privacy policy
policy.returns          — return/refund policy
pricing.general         — pricing information
catalog.product         — product details
catalog.inventory       — stock/availability
content.article         — blog posts, guides
people.exec             — leadership, staff
opinions.review         — customer reviews

6. Authority Rules

Authority Rules define which surfaces are authoritative for which claim types, and how conflicts are resolved.

Conflict Resolution

When two surfaces disagree on a claim, the rule's conflict_policy determines the outcome:

• prefer_authoritative — the primary surface wins
• prefer_recent — the most recently updated surface wins
• flag_for_review — the conflict is escalated for human review

Tie-breakers are applied in order: surface role, proof strength, then recency.

7. The .well-known/entity.json Endpoint

The Authority Graph is served as a JSON document at /.well-known/entity.json. The schema:

{
  "schema_version": "aitwire.authority.graph.v1",
  "publisher": { "tenant": "<tenant_id>" },
  "subject": {
    "tenant": "<tenant_id>",
    "canonical_domains": ["example.com"]
  },
  "issued_at": <unix_timestamp>,
  "validity_window": {
    "not_before": <unix_timestamp>,
    "not_after": <unix_timestamp>
  },
  "endpoints": {
    "authority": "/.well-known/entity.json",
    "proofs": "/.well-known/authority.proofs.json",
    "governance": "/.well-known/ai-entity.json"
  },
  "surfaces": [ ... ],
  "proofs_index": [ ... ],
  "authority_rules": [ ... ],
  "claim_types": [ ... ],
  "feeds": [ ... ]
}

8. Feed & Schema Endpoints

The Authority Graph declares structured data feeds for each content type. Feeds are JSON endpoints that return paginated, versioned records:

• /feeds/products.json — product catalog
• /feeds/articles.json — published articles and guides
• /feeds/locations.json — physical locations
• /feeds/services.json — service offerings
• /feeds/policies.json — business policies
• /feeds/pages.json — canonical page inventory

Schema endpoints provide JSON-LD representations of individual entities, enabling AI Systems to extract structured data deterministically.

9. Versioning & Validity

Every Authority Graph has a validity_window that declares when the document should be considered fresh. AI Systems should re-fetch the document after not_after.

The document includes an ETag and Last-Modified header for conditional requests. Clients should use If-None-Match or If-Modified-Since to minimize bandwidth.

Entity versions are content-addressed: a SHA-256 hash of the entity's canonical fields determines the version. Unchanged content does not increment the version counter.

10. Revocation

Surfaces and proofs can be revoked at any time. Revoked surfaces are marked with a revoked_at timestamp and excluded from active Authority Graph resolution. AI Systems should check for revocations before citing data from a surface.

11. Governance & Drift

AITWIRE monitors surfaces for drift — when the data on a surface diverges from the authoritative record. Drift events are published at /.well-known/events/drift.json.

Governance thresholds define when drift becomes actionable. When a threshold is breached, enforcement actions can be triggered automatically (for Business+ plans) or flagged for manual review.

The governance pipeline integrates with the Cloudflare Edge Enhancement layer, enabling real-time header injection and crawler policy enforcement on tenant domains.