Datapoint Model

A reusable datapoint should be treated as governed disclosure evidence, not just extracted text. The datapoint needs enough context for CDP to know which discloser it belongs to, what source supports it, which standards it may help answer, and whether it is safe to reuse.

Suggested Datapoint Shape

Field	Purpose
Discloser profile ID	Connects the datapoint to the canonical organization or place.
Source evidence	Links back to uploaded documents, prior disclosures, CRM records, or approved internal sources.
Reporting period	Prevents stale values from being reused accidentally.
CDP question reference	Shows where the datapoint can support CDP disclosure.
Framework alignment tags	Shows external standards or frameworks the datapoint may help address.
Value and unit	Captures numeric, boolean, categorical, or narrative content in a structured way.
Confidence and extraction method	Separates directly supplied, human-confirmed, AI-extracted, and inferred values.
Review status	Tracks imported, extracted, suggested, user-confirmed, reviewer-approved, expired, or superseded states.
Reuse permissions	Defines where the datapoint can be reused and whether additional consent or review is required.
Caveats	Records standard-specific limitations, data-quality issues, or required expert review.

Review States

State	Meaning
Imported	A value or evidence item has been brought into the workflow but not interpreted.
Extracted	AI or automation has extracted a candidate datapoint from source evidence.
Suggested	The datapoint has been used to support a draft disclosure response.
User-confirmed	A user has accepted that the datapoint is relevant and accurate enough for the workflow.
Reviewer-approved	A qualified reviewer has approved the datapoint for reuse in defined contexts.
Expired	The datapoint should no longer be reused because its reporting period or validity window has passed.
Superseded	A newer datapoint or corrected source has replaced it.

Reuse Boundaries

A mapped datapoint should not be reused across entities, reporting years, subsidiaries, or jurisdictions unless the profile, boundary, source, and permission context supports that reuse.

The model needs to distinguish:

• organization-level evidence from subsidiary-level evidence;
• city, state, region, country, and corporate reporting boundaries;
• current reporting-year values from older source material;
• directly supplied values from inferred or AI-extracted values;
• CDP questionnaire support from external standards or regulatory compliance claims.