Skip to content
← All work

Case study / BrandArmor

Evidence for a human review, not an automatic accusation.

Evidence-first review of suspicious skincare marketplace listings. Builds product baselines and prioritizes human review, not automatic counterfeit detection.

Sign-in required for the pilot workspace. Source is public.

TypeScriptAI

The problem

Counterfeit-listing review is a high-stakes decision. Missing evidence must not become a confident accusation.

The approach

  1. Separate evidence collection and risk assessment from the human review decision.
  2. Document delivered scope and known limits rather than promise automatic counterfeit detection.

What the evidence shows

A public repository and sign-in-protected pilot explain the evidence-first workflow. Reported repository checks are evidence of the build process, not independently measured detection accuracy.

Known limits

Sign-in is required. No customer, reviewed-case count, accuracy metric, or unverified competition placement is claimed.

Inspect the evidence

Have a related problem?

Request a scoped quote ↗