Case Studies

Major Mobile App Store: Malicious App Detection at Platform Scale

Major mobile app store distributing over 3M apps.

Global Design Platform: Employee Impersonation Detection on LinkedIn

Collaborative design platform used by 95% of the Fortune 500.

Large Payments Platform: Detection of Illegal Gambling Accounts

Major Indonesian payments platform serving 110M+ users.

Regulated Neobank: Automated Cross-Channel Abuse Detection

Consumer neobank serving 10M+ customers.

Major Mobile App Store: Malicious App Detection at Platform Scale

Context

Major mobile app store distributing over 3M applications globally. The platform sought to protect users from malicious apps impersonating trusted brands. This case study reflects detection and enforcement activity over a 12-month period.

Challenge

Advanced visual analysis was required to detect adversarial impersonation at platform scale.

Outcome

  • 2,400+ impersonation apps submitted
  • 72% action rate (Suspend / Block / Warning)
  • Impersonation detected across 500+ brands
  • Flagged scam apps representing 50M+ cumulative installs
  • Documented a 30% recurrence rate via version-based evasion

System Capabilities Demonstrated

  • Advanced visual detection of brand-color mimicry and splashscreen manipulation
  • Identification of repeat offenders and published/unpublished state evasion
  • Reverse-engineering of mobile app code to uncover hidden trading functionality
  • Detection of geolocation-triggered malicious behavior

Large Payments Platform: Detection of Illegal Gambling Accounts

Context

Indonesian payments platform serving 110M users. Indonesia has strict prohibitions on online gambling. The platform sought to identify payment accounts used to receive illegal gambling payments. Findings are based on detection data from a 4-month period.

Challenge

Advanced web application execution capabilities required to confirm misuse through registration and login flows and extract payment identifiers from authenticated states.

Outcome

  • 800+ payment identifiers identified
  • Phone and bank account numbers extracted
  • QRIS codes extracted

System Capabilities Demonstrated

  • Visual analysis to identify gambling domains advertising payment acceptance
  • Automated execution of registration and login flows
  • Extraction of payment identifiers from authenticated states

Global Design Platform: Employee Impersonation Detection on LinkedIn

Context

Collaborative design platform used by 95% of the Fortune 500. The company sought to identify and takedown employee impersonation on LinkedIn. Findings are based on detection data from a 12-month period.

Challenge

Multimodal profile analysis was required to identify fraudulent LinkedIn profiles.

Outcome

  • 10,000+ fraudulent LinkedIn profiles identified
  • Malicious profiles with verified badges flagged
  • Surface reduced to legitimate affiliation baseline
  • Broader community and freelance profiles preserved

System Capabilities Demonstrated

  • Multimodal analysis of profile text, experience history, and visual identity
  • Profile classification and disambiguation at scale
  • Continuous monitoring for fraudulent profiles

Regulated Neobank: Automated Cross-Channel Abuse Detection

Context

Consumer neobank serving 10M+ customers, regulated by the Reserve Bank of India (RBI). The platform faced daily phishing, impersonation, and credential theft attacks. Findings are based on detection and enforcement data from a 12-month period.

Challenge

Automated cross-channel detection was required to identify and drive takedown abuse targeting web, mobile, and social product surfaces.

Outcome

  • 80% reduction in active scams within three months
  • Identified over 1000 scams across websites, social media, app stores, and the dark web
  • 50% increase in scam detection coverage vs. prior solution
  • Proactive blocking of compromised customer payment cards

System Capabilities Demonstrated

  • Automated cross-channel detection 
  • Visual analysis of scams
  • Continuous monitoring for leaked cards and employee credentials
  • Structured extraction of scam intelligence to strengthen KYC control