Clear about the basis.
Clear about the limits.
The approach behind evidence that people can understand, question and use.
Start with the identity and timing.
Evidence needs an attributable source: the device, its assignment to a property and room, and when the measurement was made. Changes in assignment or equipment must remain understandable when looking back.
Make data quality visible.
Readings need context about units, plausibility, reporting interval, coverage and missingness. A disconnected device or incomplete period must not be presented as a reassuring result.
Use a reference that applies.
References may come from government, regulators, technical standards, scientific literature, manufacturers or the customer’s operating rules. They serve different purposes and must remain clearly labelled.
The source, version, measurement, unit, time basis and applicability matter. A published guideline is not automatically a statutory limit.
Keep comparisons reviewable.
Where the evidence and reference support it, calculations can describe departure, duration, persistence, recurrence and trend. The basis should be sufficiently clear for the result to be checked and reproduced.
Generative AI is not the authority for material evidence evaluation, hazard classification or legal conclusions. Any explanatory or summarising assistance remains separate from the underlying measurements and calculations.
Keep judgement with responsible people.
HiveCheck does not assess or decide risk, diagnose a building, determine an HHSRS category, decide whether Awaab’s Law has been triggered, establish causation, prescribe a repair or certify legal compliance.
Evidence supports resident reports, inspections and professional judgement. Sensors and connectivity can fail or be delayed, and every deployment needs an appropriate response process.
Read the original sources.
Use the current source material for its intended purpose and jurisdiction.