We measure whether the system helps — not who a person is.
Smart evaluates products through task success, reliability, correction, human handoff and evidence of useful outcomes. Measurement should work like a clear dashboard, not a hidden scorecard: it should show whether the journey works and where the product needs attention without turning the user into the product.
What we measure: whether the intended user can complete the defined task; whether the system behaves consistently and recovers safely when a dependency fails; whether important output can be checked and how often it must be corrected; whether the system recognises its boundary and transfers context to a person; whether the workflow reduces avoidable steps; whether people can navigate and understand across supported devices, languages and assistive technologies; and whether the information collected stays proportionate to its purpose.
What we refuse to measure or infer: Smart does not use routine browsing, biometric or demographic signals to rank a person’s worth, intelligence, wealth, personality or entitlement to facts and respect. Product improvement focuses on the journey and the system — not a hidden judgement of the individual.
Evidence hierarchy, strongest first: working behaviour that can be repeated or verified; controlled test evidence with stated conditions; operational measurement with a defined period, sample and limitation; voluntary user feedback; and finally clearly labelled roadmap intention, which is never presented as current evidence.
On this website: product state is read from one canonical source; charts show the measure, period and source; small or unreliable samples stay unpublished or explicitly limited; testimonials, ratings and outcome claims require real approved evidence; and every product page separates what works now from what remains under validation.