Home/Knowledge/AGENT CONTENT NETWORKS
AGENT CONTENT NETWORKSMarch 202611 min read

How Agent Content Networks Work: Register, Discover, Cite

A complete technical walkthrough of the three-phase lifecycle: how data enters the network, how AI agents find it, and how every interaction generates attribution intelligence.

KEY TAKEAWAYS
—Registration involves domain verification (DNS TXT), signal intelligence (fraud detection), and Ed25519 keypair issuance
—Discovery uses weighted text search: business names 10x, categories 5x, service areas 3x, facts 2x
—Every query generates a citation record: what was asked, which businesses returned, which AI agent asked
—Content objects have ~30 typed fields — structured data AI agents can query directly
—The MCP server enables any compatible AI platform to connect to the ACN as a tool

Phase 1: Register

Data enters the Agent Content Network through two paths: automated ingestion and publisher self-registration.

Automated Ingestion

The ACN's seed pipeline ingests business data from public sources. Each business becomes a typed content object with structured fields for name, category, address, phone, hours, services, facts, and more. All ingested data is marked unverified.

Publisher Self-Registration

  1. Register — POST email, domain, and business name. Layer 2 signal intelligence runs immediately.
  2. Verify — Place a DNS TXT record containing the challenge token. Proves domain ownership.
  3. Receive Credentials — Ed25519 keypair generated. Private key returned once, never stored. API key issued.

Phase 2: Discover

AI agents discover businesses by querying the ACN through the MCP server or REST API.

Search Weight Distribution

Name: 10x | Category: 5x | Service Area: 3x | Facts: 2x | Description: 1x

Results are sorted by three signals: verified businesses first, then text relevance, then completeness score. This incentivizes businesses to both verify and complete their profiles.

Phase 3: Cite

Every query generates a citation record logging: query text, returned businesses, agent identifier, fields accessed, and timestamp.

For verified businesses, this powers citation intelligence: how many times recommended this week, which AI platforms cite you most, what questions lead to your business, and competitor comparison.

Google tells you how many people searched for you. The ACN tells you how many AI agents recommended you.

Get on the network

Search the ACN to see if your business is already indexed — then claim your listing.

Search the ACN →
FREQUENTLY ASKED QUESTIONS
How does the ACN prevent fake listings?
Three-layer trust: DNS verification proves ownership. Signal intelligence detects fraud. Ed25519 chain of custody ensures only verified owners can update data.
What is the MCP server?
MCP (Model Context Protocol) is a standard allowing AI platforms to connect to external tools. The ACN's MCP server exposes search, retrieval, facts, Q&A, and verification tools.
Can I add custom facts?
Yes. Verified publishers add facts, Q&A pairs, services through signed content updates. Every update must be cryptographically signed.
How fast is ACN search?
Typical query execution is under 50ms. The weighted text index is optimized for AI agent query patterns.

Is your business visible to AI?

Find out in 30 seconds. GeoProof scans how AI platforms see your business right now.

Run Free AI Audit →
CONTINUE READING
AGENT CONTENT NETWORKS
What Is an Agent Content Network? The Definitive Guide
12 min read
AGENT CONTENT NETWORKS
Agent Content Network vs Search Engine
10 min read
AGENT CONTENT NETWORKS
Citation Tracking
8 min read