Landscape · deep-dive · Schema.org
agentic-first vs Schema.org.
Schema.org is the right answer for SEO and Knowledge-Panel discovery. It is not the right answer for agent-grade diligence, and it was never designed to be. Ship both. They describe different things and don't conflict.
What Schema.org is
Schema.org is a
shared vocabulary maintained by the Schema.org community
(Google, Microsoft, Yahoo, Yandex). The core
Organization and Person types let you
embed structured data on a homepage as JSON-LD inside a
<script type="application/ld+json"> block.
Search engines parse it for rich results, Knowledge Panels, and
sitelinks. Most current LLM crawlers also read it.
It does what it does very well. Adoption is universal, the rendering pipelines exist, and the vocabulary is stable enough to bet a website on. If you only ship one structured-data file, ship Schema.org.
Side-by-side
| Question | Schema.org | agentic-first |
|---|---|---|
| Primary consumer | Search engines (Google, Bing) and general LLM crawlers | Agents doing diligence; the directory; investor / sourcing tooling |
| Where it lives | <script type="application/ld+json"> in your homepage HTML |
/.well-known/agentic-profile.json at a known path |
| Identity anchor | Free-form name, url, sameAs |
Statutory IDs: company.registry (Companies House etc.) + company.lei (GLEIF), verified by the directory |
| Banded financials | Not modelled | Banded funding, metrics.revenue_band, metrics.growth_band - FCA-aware |
| Stage / fundraise state | Not modelled | Enum stage.current (Pre-Seed → Public → Acquired); banded last-round detail |
| Contact-channel preference | contactPoint array, free-form |
Enum contact.preferred_channel (email / form / private-mcp / none) |
| Evidence per claim | Not modelled | evidence[] with JSON-Pointer in supports back to the field |
| Protected-tier surface | Not modelled | contact.private_mcp pointer; OAuth-scoped MCP behind your own auth |
| Versioning | Schema.org is loosely versioned by date; consumers tolerate drift | SemVer in schema_version; immutable schema URLs at /schemas/{name}-{version}.json |
| Validation feedback | Google's Rich Results Test, Schema Markup Validator | submit_website MCP tool returns structured errors + warnings + score-input vector |
When to use which
- Use Schema.org for everything a general search engine or non-specialist LLM crawler needs - your name, logo, address, social profiles, opening hours, simple
contactPoint. If your goal is "show up in Google with a Knowledge Panel", this is the file. - Use agentic-first for everything a serious agent needs once it's decided your company is a candidate - jurisdiction, registry IDs, stage, banded financials, evidence, the contact channel you actually want used, and the pointer to your protected-tier MCP. If your goal is "show up correctly when an investor agent runs a sourcing brief", this is the file.
How they compose on the same homepage
Both files target the same domain root, so the recommended setup is to ship them together:
<!-- index.html - Schema.org for general discovery -->
<script type="application/ld+json">
{
"@context": "https://schema.org",
"@type": "Organization",
"name": "Acme Robotics",
"url": "https://acme-robotics.example",
"logo": "https://acme-robotics.example/logo.svg",
"sameAs": [
"https://www.linkedin.com/company/acme-robotics",
"https://github.com/acme-robotics"
]
}
</script>
# /.well-known/agentic-profile.json - agent-grade diligence
{
"schema_version": "0.1.0",
"updated_at": "2026-04-19T00:00:00Z",
"profile_kind": "company",
"tier": "public",
"company": {
"name": "Acme Robotics",
"website": "https://acme-robotics.example",
"jurisdiction": "GB",
"registry": { "type": "companies-house", "id": "12345678",
"url": "https://find-and-update.company-information.service.gov.uk/company/12345678" }
},
"stage": { "current": "Seed" },
"funding": { "total_raised_band": "1m-5m", "currency": "GBP" },
"team": { "headcount_band": "11-50" },
"metrics": { "revenue_band": "100k-500k", "as_of": "2026-03-31" },
"contact": { "preferred_channel": "form",
"form_url": "https://acme-robotics.example/contact",
"private_mcp":"https://private-mcp.acme-robotics.example/mcp" }
}
The two are complementary - Schema.org is read by the general web, agentic-first is read by directories, agents, and the occasional human looking at JSON. Neither file overrides the other; they describe different audiences.
The honest summary
We're not trying to compete with Schema.org and we'd be silly to. The vocabulary is older, broader, more adopted, and aimed at a different reader. What Schema.org doesn't carry - banded financials, registry-anchored identity, fundraise stage, protected-tier pointers, evidence per claim - agentic-first does. If you're publishing structured data about your company at all, ship both.