blueproof.ai is ScopeLimited's measurement-and-reporting product for connected water fixtures. It keeps source-side activity, the modeled baseline, method, assumptions, exceptions, and calculation together so a reporting team or independent reviewer can test how the result was produced.
Connected fixtures report activity and water-use information through the manufacturer's platform. blueproof.ai preserves that source-side record with fixture identity and product context rather than inferring one project from a whole-building utility bill.
The baseline is a counterfactual supported by pre-project specifications, building records, code, and disclosed rules. It is labeled as modeled and kept distinct from post-installation fixture evidence.
The method calculates avoided withdrawal from the modeled without-project case and device-reported project use, with conservative treatment for gaps and anomalous activity. The method is mapped to WRI's voluntary VWBA 2.0 guidance.
The record includes source-reading hashes, baseline basis, method version, assumptions, exclusions, data-availability decisions, and calculation. Signed, hash-linked records make later changes visible.
A reviewer can inspect the source evidence, challenge the baseline and assumptions, and rerun the calculation. blueproof.ai prepares the claim-support record; an independent practitioner decides whether to provide assurance.
The record is not a certificate or assurance opinion. It is the material needed to test the project result: source-side fixture evidence, the separately modeled baseline, method, assumptions, exceptions, and calculation.
Illustrative of our engine's output — sample data, not live.
Each signed record includes the previous record's hash. Change a past input or result and the digest no longer reconciles with the stored signature and chain. That protects record integrity; it does not replace review of the method, baseline, or source evidence.
Each record's prev is the previous record's hash, highlighted where they match. Batches can be Merkle-rooted and Ed25519-signed; a pilot determines the custody or public-anchoring pattern appropriate to the reporting use.
Fixture activity and device-reported use describe the project after installation. The baseline remains a modeled comparison supported by separate evidence. The record labels both clearly so a reviewer can see where observation ends and modeling begins.
Illustrative figures — one fixture, one day — to show the structure of the calculation, not a reported result.
Fixture activity, reported use, identity, and product context ingested through the manufacturer's platform.
The source evidence, modeled baseline, method, assumptions, exceptions, and calculation remain together.
An internal reviewer or independent practitioner tests the evidence and decides whether the result supports the intended use.
We can walk through the sample receipt, data-flow diagram, VWBA 2.0 mapping, verifier export shape, and claims-language guardrails.