Governance

Open governance for an open standard

Charter

Mission Statement

The AI Discovery Standard organization exists to develop, maintain, and promote an open specification for structured, machine-readable organizational identity at /.well-known/ai — enabling every website to communicate directly with AI agents in a verifiable, vendor-neutral format.

Principles

  1. Open specification. The standard is published under CC-BY-4.0. Anyone may implement it without permission, licensing fees, or membership requirements.
  2. Multi-vendor governance. No single company controls the specification. The Steering Committee represents diverse stakeholders.
  3. Interoperability first. Changes to the specification prioritize backward compatibility and interoperability between implementations.
  4. Minimal viable spec. The specification defines what is necessary for interoperability and no more. Implementers are free to extend beyond the spec.
  5. Implementation-driven. Specification changes must be demonstrated in at least one reference implementation before advancing to final status.

Steering Committee

The Steering Committee provides strategic direction, approves specification releases, and resolves disputes. Seats are allocated to ensure representation from technology providers, enterprise adopters, and the broader community.

Steven Sprague

Rootz Corp · Chair

Standard author. Brings 25 years of standards work from the Trusted Computing Group. Responsible for specification direction and technical coherence.

Adam Reiser

Made in USA Inc.

Enterprise and government adoption. Represents the certification and managed-service implementation perspective.

Founding Member Seat

Technology Provider

Reserved for a major technology company implementing the standard in their platform or infrastructure.

Founding Member Seat

Enterprise Adopter

Reserved for an enterprise organization deploying the standard across their web properties.

Founding Member Seat

AI Platform

Reserved for an AI platform or agent framework that consumes AI Discovery data.

Community Seat

Open Working Group

Elected by active working group participants. No membership requirement for candidacy.

Working Groups

Working groups are where the specification gets developed. Participation is open to anyone — no membership required. This is intentional: engineers should be able to contribute before their legal department approves a membership agreement.

Core Specification WG

Active

Maintains the core specification, reviews proposed changes, and ensures backward compatibility.

  • Schema definition and validation
  • Required vs. optional field decisions
  • Versioning and migration paths
  • IANA registration for /.well-known/ai

Scoring & Certification WG

Forming

Defines how implementations are scored and how scanners normalize their results for comparability.

  • Score normalization model
  • Tier weighting (Discovery, Knowledge, Feed)
  • Grade boundaries (A/B/C/D/F)
  • Scanner interoperability testing

Signing & Attestation WG

Forming

Develops the cryptographic signing model, attestation manifests, and verification protocols.

  • Signature block specification
  • Certificate anchoring patterns
  • Post-quantum readiness
  • Multi-signer governance

Implementations WG

Planned

Supports implementers, maintains reference code, develops test suites, and validates interoperability.

  • WordPress plugin reference
  • Static site generators
  • CMS integrations
  • Conformance test suite

Specification Process

Changes to the specification follow a structured process modeled on W3C and IETF practices.

1

Proposal

Anyone may submit a proposal to a working group. Proposals describe the problem, proposed solution, and impact on existing implementations.

2

Working Draft

The working group reviews and refines the proposal into a working draft. This stage involves discussion, iteration, and consensus-building.

3

Reference Implementation

At least one reference implementation must demonstrate the proposed change. "Code talks" — a feature that can't be implemented is a feature that shouldn't be specified.

4

Candidate Recommendation

The Steering Committee reviews the working draft and reference implementation. If approved, the change enters a public review period.

5

Final Specification

After public review and any necessary revisions, the Steering Committee votes to incorporate the change into the next specification release.

Intellectual Property Policy

The Standard is Free

The AI Discovery Standard specification is published under CC-BY-4.0. This means:

Anyone may implement the specification, build products based on it, create derivative specifications, and use it commercially — with attribution to the AI Discovery Standard project.

No patent encumbrance. Working group participants agree to license any essential claims under royalty-free terms to all implementers of the specification.

Contribution Agreement

By participating in a working group, contributors agree that their contributions to the specification may be incorporated under the CC-BY-4.0 license. This is a lightweight agreement — no CLA required for participating in discussions, reviewing drafts, or suggesting changes.

What Remains Proprietary

While the specification is open, implementations may include proprietary extensions, enterprise features, and commercial tooling. The boundary is clear:

  • Open: The specification, required fields, JSON schema, endpoint paths, signing format, scoring categories
  • May be proprietary: Scoring algorithms, enterprise dashboards, managed services, attestation infrastructure, compliance mapping tools

This mirrors the pattern used by SSL/TLS (open spec, commercial certificate authorities), HTTP (open spec, commercial web servers), and HTML (open spec, commercial browsers).