Evidence first
Every buy, wait or watch note should be tied to visible signals such as price history, stock state, seller availability and review context.
AI transparency
Scoutly AI explains signals, uncertainty and limits.
Scoutly AI is designed as decision support for product research and commerce intelligence. It does not present shopping outcomes as guaranteed predictions.
Principles
Recommendations should be explainable and auditable.
No false certainty
The product avoids guaranteed language. It shows risk labels, signal strength and uncertainty where data is incomplete.
User control
Users decide whether to act. Scoutly AI provides context, alerts and explanations, not final purchasing authority.
Model evaluation
Claude-generated summaries should be checked against structured source signals to reduce overstatement and unsupported claims.
Claude usage boundary
Claude is planned for structured summaries, product reasoning, risk labels, decision notes and alert explanations. Source data and business logic remain part of the Scoutly AI product layer.
What users should see
Users should see why a product is marked buy, wait or watch, which signals influenced the note and where the system is uncertain.
What Scoutly AI avoids
Scoutly AI avoids presenting generated output as financial advice, guaranteed savings, guaranteed availability or guaranteed seller reliability.