Define once. Publish once. Let every compatible system start from the same source.
These open public community standards define predictable structures and discovery locations for every organization's standards. An organization publishes its authoritative information using the applicable schema. AI systems, applications, vendors, and people can then discover and retrieve the specific information relevant to their task without requiring the organization to rebuild its context for each tool.
1. Define authoritative information
The organization decides what is authoritative: which product names are correct, which claims are approved, what a term means, how a data object is defined. This is a decision about ownership, not formatting.
2. Structure it using the applicable SchemaFirst schema
The information is expressed in the schema that fits it—Brand Schema for brand and identity, the GTM Data Schema for go-to-market data objects. Structure is what makes the same source readable by many different consumers.
3. Publish it at a stable canonical location
The structured source is published at a predictable, canonical location the organization controls. Public context goes to a well-known path on your own domain—for example, your public brand at www.yourbrand.com/well-known.brand-schema. Proprietary context—like customer segments—stays behind your walls, published to your wiki or SharePoint in the same standard format.
Stability of location and data format is what makes the standard scalable: any system that knows the convention can find the source and read it without a custom integration.
4. Discover the available standards
A tool, agent, or person can find which standards an organization has published without being told out of band. Discovery uses conventional locations and references rather than private, per-vendor configuration.
Because every vendor reads from the same public schema, mapping happens once against the standard—not once per tool. Data mapping becomes a one-time exercise instead of a recurring integration cost that repeats with every new system you adopt.
5. Retrieve only the relevant information
Each consumer pulls the specific fields it needs. A design system retrieves visual identity; a sales agent retrieves product terminology and approved claims. Retrieval is scoped to the task, not to the whole source.
6. Apply it within the tool or workflow
The retrieved context is used inside the consuming tool—generating copy, mapping records, checking a design. SchemaFirst does not dictate how the tool works internally; it standardizes the information the tool starts from.
7. Validate the output against the source and version
Because outputs reference a specific source and version, they can be checked against it. Validation answers a concrete question: does this output match the standard that governed it?
8. Update the canonical source as the organization changes
When something changes, the organization updates the canonical source once. New retrievals reflect the current truth, while historical outputs remain tied to the version that produced them.
The model at a glance
Organization
↓ publishes
Canonical SchemaFirst sources
↓ discovered and retrieved by
AI tools · applications · vendors · teams
↓ produce
Content · data · designs · actions
↓ checked against
Applicable standard and versionNot every consumer needs every standard
A design system may consume visual identity categories while a sales agent consumes product terminology and approved claims. The same published sources serve many consumers, each retrieving a different, relevant subset. Publishing is universal; consumption is selective.
Governing standards
Explore the Standards
Explore the StandardsLast reviewed · Maintained by the SchemaFirst Standards Working Group