Technical trust

What the evidence includes and what it can establish.

A reviewer needs the conclusion and a way to examine its basis. Headroom's exported-data workflow produces a report with capacity states, reasons, data limitations, and supporting evidence and replay materials.

Test a Power Claim See a Sample Decision

What a reviewer receives

A screening report with capacity limits, principal constraints, reasons, and next checks.
A record of the supplied inputs, evaluation scope, protected obligations, and assumptions.
Findings on missing, stale, incomplete, or conflicting data.
An evidence-file inventory, integrity checks, and replay or verification materials for the agreed workflow.

Where the evidence starts

The evaluation relies on authorized exported records and the assumptions supplied for the question. Data-quality checks can flag limitations and reduce confidence; they cannot recover facts that were never measured or supplied.

Protecting reserve, charging, or other higher-priority obligations can limit the proposed load without meaning the site itself failed. The report separates those outcomes.

Three different checks

Reproducibility, integrity, and measurement accuracy.

Can the result be reproduced?

Replay the supported evaluation with the same inputs, assumptions, and policy version, then compare the resulting states and reasons. This checks the computation within that scope.

Has the saved package changed?

Compare saved files with a trusted reference manifest. Matching hashes can show that those files have not changed since that reference was recorded.

Were the measurements accurate?

That requires source verification, such as checking meters, calibration, and collection methods. Read-only operation and matching hashes do not prove the physical accuracy of supplied data.

Current status

Exported-data evaluation within a defined scope.

The current evidence-pilot release supports exported-data intake, data-quality assessment, ranked obligations, capacity states, reports, manifests, and replay materials. Its demonstrations use test fixtures and do not establish customer validation. Live deployment and specific integrations are not claimed; recurring and embedded use remain evaluation topics.

Inspect the example's supporting evidence or discuss a partner evaluation.