Trust · transparency · accountability

How Workforce Wonkery earns trust

Workforce Wonkery is an independent educational resource. It uses AI-assisted research, drafting, organization, and source checking. Policy briefs do not receive human legal or compliance review. The goal is not to replace official guidance, but to help readers understand it while making the limits of the analysis visible.

What happens before a policy brief can be public

1. Primary-source capture. The applicable official source must be identified and linked whenever available.
2. Structured extraction. An AI pass extracts dates, status, issuing authority, requirements, thresholds, deadlines, and other consequential facts from the primary source.
3. Independent cross-check. A separate AI review compares the draft brief against the primary source and looks for contradiction, overstatement, omission, ambiguity, or unsupported interpretation.
4. Supersession check. The workflow looks for later guidance, rescission, amendment, replacement, or other signs that the source may no longer control.
5. Uncertainty gate. Material conflicts or unresolved ambiguity block automatic publication or must be stated conspicuously.
6. Trust record. Public briefs use a two-level trust display: a compact source-check signal near the top, followed by a full Source + Trust Record at the bottom. The full record identifies the primary authority, current status, source-check date, what was verified, important limitations, and the fact that the brief is AI-assisted and does not receive human legal or compliance review.

What “checked against primary source” means

It means an automated comparison was performed against the cited official material. It does not mean the brief was approved by an agency, reviewed by counsel, certified for compliance, or guaranteed to be complete or correct.

High-consequence topics

For eligibility, fiscal controls, procurement, contracting, monitoring, civil rights, participant costs, sanctions, deadlines, personnel, and similar matters, the site should avoid presenting editorial interpretation as an instruction to act. These pages should emphasize what the source appears to require, questions to resolve before acting, and what should be confirmed locally.

When the answer is uncertain

If authoritative sources are ambiguous, incomplete, conflicting, pending, or appear to have changed, Workforce Wonkery should say so. A page may remain private rather than present an uncertain interpretation as settled.

Workforce learning trust model

Workforce Wonkery learning resources use a tiered trust standard so foundational learning is not burdened with the same warnings as a high-consequence decision aid.

• Learning and orientation. Workforce Foundations and other foundational pages explain the system in plain language. They disclose AI assistance and remind readers that official sources control when requirements, rights, funding, or authority are at issue.
• Operational guidance. Workforce in Practice, reference tools, and decision aids disclose that they are AI-assisted and not human legal/compliance reviewed. Claims about authority, funding, performance, and compliance should be source-traceable and checked against authoritative material.
• High-consequence Quickstarts. Eligibility, priority of service, procurement, fiscal, monitoring, contracts, participant payments, civil rights, data corrections, and similar playbooks are framed as orientation, not instruction. They should help readers identify governing sources, questions, documentation, and escalation points—not substitute for the actual decision authority.

Across all three tiers, the site aims to distinguish what is required, what California adds, what a local area can decide, and what is presented as good practice. When that boundary is unclear, the page should say so and direct the reader back to the controlling source.

Intelligence and labor-market methodology

Workforce Wonkery intelligence has a different trust problem from policy summaries: the risk is not only getting a fact wrong, but making interpretation, causation, selection, or a scenario look more certain than the evidence supports.

• Separate evidence from interpretation. Data and documented events should be distinguishable from the Wonkery Read, scenarios, and operational considerations.
• Prefer primary data and official sources. Government datasets, official records, and first-party public documents are preferred for factual claims. Reputable secondary reporting may add context and should be identifiable as reporting rather than authority.
• Show the data period, not just the publication date. Labor-market claims should identify the relevant reference period, geography, and whether figures are seasonally adjusted, preliminary, revised, or estimated when material.
• Label estimates. Derived figures should be described as estimates and, when material, explain the source inputs or method used to construct them.
• Avoid causal overreach. Timing or correlation alone is not treated as proof that one event caused another. The analysis should use appropriately conditional language when causal evidence is limited.
• Scenarios are not forecasts. Forward-looking policy, budget, or economic analysis should identify the condition being modeled and avoid unsupported probability claims.
• Preserve conflicting evidence. If indicators point in different directions, the analysis should show the tension rather than force a single story.
• Local examples are editorial selections. California Workforce Board Examples and other practice scans are not rankings, endorsements, or comprehensive surveys. Publication date and action date should not be conflated.
• Material revisions should be visible. Later data, corrections, or changed circumstances that materially alter the interpretation should be reflected in the update record.

Intelligence products are AI-assisted and do not receive human legal, compliance, or editorial review. Readers should inspect cited sources and underlying data before relying on a conclusion for a consequential decision.

Visuals and media provenance

Some Workforce Wonkery illustrations and decorative graphics are AI-generated. Generated visuals are used as editorial illustration or design, not as documentary evidence that a depicted person, place, meeting, event, or scene actually occurred. A visual should not be used to imply factual proof that is not supported by the cited sources.

Third-party photographs, screenshots, charts, maps, logos, and other external media should be attributed or used under an appropriate permission, license, public-domain status, or other lawful basis. When an external visual is itself evidence, the surrounding content should identify the source and context rather than relying on the image alone.

Attribution and use of outside material

Workforce Wonkery should synthesize facts and ideas without copying another source’s distinctive expression. External reporting, research, charts, images, and examples should be attributed where used. Quotations should be limited and purposeful. When a source provides the underlying authority or evidence, readers should be able to reach that source directly whenever practicable.

Material relationships and conflicts

The site does not need to publish a contributor’s full professional biography to be transparent. But when a material professional, financial, organizational, or project relationship bears directly on a program, organization, initiative, or example being discussed, that relationship should be disclosed in the relevant content when it could reasonably affect how a reader evaluates the analysis.

Accessibility

The site aims toward WCAG 2.2 Level AA where practicable, but does not claim full conformance without an independent audit. Accessibility barriers can be reported through the challenge process. Read the accessibility statement →

Corrections and challenges

Readers can challenge a factual statement, source, policy status, data point, interpretation, characterization, or accessibility barrier. Material corrections should be documented rather than silently erased from the record. Challenge or report a problem → · Corrections & Revisions Ledger →

Independence

Unless expressly stated otherwise, analysis and editorial choices on this site are independent and do not represent the official position of any government agency, workforce development board, employer, partner organization, or other affiliated entity. Material relationships that directly bear on a covered organization, project, or example should be disclosed when they could reasonably affect how a reader evaluates the analysis.

Last updated September 18, 2026. See also Editorial Standards & Disclaimer and Privacy Policy.