Structured Data

Why structured data matters for AI agents

AI agents parse structured data before or alongside raw HTML when evaluating a page. Well-formed structured data gives agents explicit, unambiguous signals about your products, organization, and capabilities. Without it, agents must infer this information from unstructured content — a process that is less accurate and often incomplete.

Structured data directly influences whether an agent recommends your product, how it describes your offering, and how confidently it presents your information to users.

JSON-LD enhancement

Inception Agents enriches your existing schema.org JSON-LD markup with additional fields that AI agents prioritize during evaluation. If you already have structured data on your pages, the platform augments it. If you do not, it generates it from your crawled content.

Enrichment happens at the edge — your source HTML is unchanged. Agent traffic receives the enhanced version; regular browser traffic sees your original markup.

Product structured data

Product pages receive enhanced Product and Offer schema with:

• name — Canonical product name.
• description — Factual, agent-optimized product description.
• pricing — Current prices, currency, billing period, and tier structure.
• availability — In stock, out of stock, pre-order, or discontinued status.
• reviews — Aggregated rating, review count, and review distribution.
• competitive positioning — Category context and differentiators, structured as additionalProperty fields.
• feature comparisons — Key features with values, formatted for direct comparison by shopping agents.

Organization structured data

Your domain-level pages receive enhanced Organization schema with:

• Company information — Legal name, founding date, headquarters location, employee count.
• Trust signals — Security certifications, compliance standards, industry memberships.
• Contact points — Support channels, sales contacts, and response time expectations.
• Social profiles — Verified links to official social media accounts.

Agent Card

The Agent Card is a JSON file served at /.well-known/agent.json that declares your site’s capabilities to agent platforms. It acts as a machine-readable manifest that tells agents what your site can do, what protocols it supports, and how to interact with it.

Example agent.json

{
  "schema_version": "1.0",
  "name": "Acme Cloud Platform",
  "description": "Cloud infrastructure for deploying and scaling web applications.",
  "url": "https://acme.dev",
  "capabilities": {
    "product_catalog": true,
    "pricing_api": true,
    "checkout": false,
    "support_chat": true
  },
  "protocols": {
    "llms_txt": "https://acme.dev/llms.txt",
    "product_feed": "https://acme.dev/feeds/products.json",
    "openapi": "https://acme.dev/.well-known/openapi.json"
  },
  "contact": {
    "sales": "sales@acme.dev",
    "support": "https://acme.dev/support"
  },
  "trust": {
    "soc2": true,
    "gdpr_compliant": true,
    "uptime_sla": "99.99%"
  }
}

Inception Agents generates and serves your Agent Card automatically based on your site configuration and detected capabilities.

Product feeds

Product feeds are machine-readable catalogs of your products, formatted for consumption by commerce-focused AI agents. These feeds include complete product listings with pricing, availability, specifications, and categorization.

Inception Agents generates product feeds in JSON and XML formats, updated automatically when your site content changes.

Commerce protocols

The platform supports integration with emerging agent commerce protocols:

Protocol support is configured per-integration. Enable the relevant protocols in your dashboard under Settings > Commerce Protocols.

How it works

After integration, the structured data pipeline operates automatically:

1. Your site is crawled and existing structured data is extracted.
2. Missing or incomplete schema is generated from page content.
3. Agent-specific enrichments are applied (competitive positioning, feature comparisons, trust signals).
4. Enhanced structured data is served to detected agent traffic at the edge.
5. The learning engine tests variations and optimizes field values over time.

Your original pages remain unmodified. Only agent-facing responses include the enriched markup.

Verification

Inspect the enriched JSON-LD served to agents by sending a request with an agent User-Agent string:

curl -s -H "User-Agent: GPTBot/1.0" https://yourdomain.com/your-page | \
  grep -o '<script type="application/ld+json">.*</script>'

Compare this with the standard response to see the enrichments applied:

curl -s https://yourdomain.com/your-page | \
  grep -o '<script type="application/ld+json">.*</script>'

You can also verify your Agent Card directly:

curl -s https://yourdomain.com/.well-known/agent.json