Sources

What this measures. Estimated aggregate U.S. dollar loss to retailers from returns classified as fraudulent (wardrobing, item swapping, counterfeit substitution, receipt fraud).

• National Retail Federation (NRF) — Retail Loss Prevention research.
• Appriss Insights — Retail returns fraud analytics.

What this measures. The share of retailer-processed returns that loss-prevention teams ultimately flag or deny as fraudulent; an upper-bound estimate across categories, with apparel and electronics at the higher end.

• Appriss Insights — Retail returns fraud analytics.
• National Retail Federation — Organized Retail Crime survey.

What this measures. The online apparel return rate (the share of items sold that come back), and the observed rise in beauty category returns driven by product contamination, opened seal submission, and counterfeit substitution.

• National Retail Federation — Retail sales and consumer trends research.
• British Retail Consortium (BRC) — Retail Crime Survey.

Every internal page the Ask AsliVerify assistant is allowed to cite inline. Each row carries a cite-<path> anchor so the in-chat source row can deep-link straight to it, and links to its focused source page at /sources/<slug>. The set mirrors ALLOWED_PATHS from the page-indexer cron, with /sources itself included for documentation.

• Home/The homepage frames the value prop and links every primary destination.
• Pricing/pricingPricing tiers, pilot terms, and the self-serve funnel are the most-cited commercial detail.
• About/aboutFounder background, the AsliVerify credibility story, and trust signals referenced in chat.
• FAQ/faqAllowlisted even though the route 404s today, so future FAQ content is citable without re-deploying the indexer.
• Demo/demoThe live demo funnel — image intake, three-check verdict, peer-shareable recap — that the assistant walks users through.
• Sources/sourcesSelf-listed for documentation; not yet indexed by the daily page-indexer cron so it is not directly retrievable in chat today.

Note: /sources is listed for documentation but is not yet indexed by the daily page-indexer cron — it ships with no chunks until the next allowlist refresh adds it. Pages not in the allowlist (such as /how-it-works, which has no route yet) are intentionally excluded.