Operator glossary
Review highlights ('Customers say')
The AI-written summary box at the top of your reviews — Amazon's model compresses hundreds of reviews into themes with sentiment tags — now the first review content most shoppers actually read.
Amazon's generative review highlights render a 'Customers say' paragraph plus clickable attribute chips (quality, value, fit, noise…) with positive/negative sentiment, synthesized from the review corpus. Many shoppers now read only this box — and Rufus draws on the same distilled themes when answering product questions.
You cannot edit it; you can only feed it. The leverage is upstream: fix the product issues that generate negative themes (the box is ruthless about repeating them), and drive review volume on strengths so positive attributes dominate the chips. Monitor it per hero ASIN — a new negative chip appearing is a VOC-grade early warning rendered in public.
The numbers
- Rendered from the review corpus: summary paragraph + attribute chips with sentiment; not seller-editable.
- Feeds and is fed by the same signals Rufus uses — public-facing VOC.
- A new negative chip on a hero ASIN = early warning; fix the cause, not the symptom.
Related terms