Book Demo

Lifecycle Pillar

Catch changes before they become problems

Monitor merchant web presence on policy schedules and surface material drift for analyst review. Websites change constantly. You need to notice before your regulator does.

How it works

1

Crawl merchant website

AURA navigates the merchant's website, captures full-page screenshots, extracts text, and records metadata across all discoverable pages.

2

Classify against rules

Content is classified using keyword rules first, then AI for ambiguous pages. Results are compared against the merchant's baseline profile.

3

Surface drift for review

Material changes are scored by severity and routed to the analyst queue. Minor changes are logged; category pivots trigger alerts.

Kenal AURA

Scans

Scan history and status

All Statuses
6 scans
ScanMerchantStatusResult
a3f8c1d2Warung SeribucompletedViolation
7b2e4d91Bunga Emas DigitalcompletedViolation
e9d1f0abToko CahayacompletedClean
1c4a8b33Dao Siam CraftscompletedClean
d5f2a7c8Lakan Bay Cofetching
f8b3e619Pho Minh Tradingqueued

Category pivot detection

Each merchant starts with a baseline content profile from their initial scan. Subsequent scans compare current content against this baseline to identify meaningful shifts: a fashion retailer adding pharmaceutical products, a restaurant pivoting to gambling content, an e-commerce store introducing counterfeit goods. The drift score quantifies how far a merchant has moved from their approved profile.

Tiered content classification

Tier 1 applies deterministic keyword and regex rules for known BRAM violations: fast and zero cost. Pages without clear rule matches go to Tier 2 AI classification. Ambiguous cases go to Tier 3 deep analysis with full-page screenshots. Costs stay proportional to classification difficulty.

Multilingual rule sets

Rule sets support English, Bahasa Malaysia, and Chinese (simplified), the primary commercial languages across Malaysian and ASEAN merchant websites. Rules can be configured per merchant segment with different detection sensitivity for high-risk versus standard categories. Custom rules can be added at the organizational level.

Evidence-grade capture

Each scan produces timestamped screenshots, extracted text, classification results, and content hashes. These artifacts form the evidence chain if a violation is later investigated or reported to a card scheme.

Frequently asked questions

Can monitoring run on different schedules by merchant segment?
Yes. Teams can define weekly, monthly, and risk-triggered monitoring by portfolio policy. High-risk segments can be scanned daily while standard merchants are checked weekly or monthly.
Is web drift enough to make final decisions?
No. Drift is one signal layer alongside investigation workflows, reputation monitoring, and other evidence. Significant drift triggers an investigation, not automatic enforcement.
What languages are supported for content classification?
English, Bahasa Malaysia, and Chinese (simplified). The keyword rules and AI classifiers are trained on content in all three languages.
How are false positives handled?
Analysts can review, validate, or dismiss any classification result. Dismissed results are logged with rationale for audit purposes.

Ready to take control of merchant risk?