Global Neo-Bank Neutralizes Credential Stuffing & Account Takeovers
A Tier-1 digital banking platform serving 12 million account holders experienced sustained high-volume credential stuffing attacks originating from residential proxy networks. Legacy WAF rules caused high false positives for legitimate mobile app users. Deploying Axiom Zero Layer 01–03 hardware attestation at the edge blocked 100% of proxy-relayed automation without imposing CAPTCHA challenges on legitimate banking customers.
RESULTS
- 847,293 automated account takeover attempts blocked in first 72 hours
- $2.3M in fraudulent transactions prevented in month 1
- 0.002% false positive rate — no legitimate customers impacted
- Detection latency: 11.2ms P99 at peak load (340K req/min)
- Time to integration: 6 hours from API key to full production deployment
Enterprise Retailer Stops High-Speed Inventory Scalper Bots
During limited-edition product drops, automated checkout bots (using Playwright and custom browser forks) depleted inventory in under 4 seconds, causing customer outrage. Axiom Zero deployed Layer 04 (Physical Display VSync Analysis) and Layer 06 (Adversarial Traps), instantly identifying headless browser instances and routing scalper traffic to synthetic GHOST checkout confirmations while preserving inventory for verified human buyers.
RESULTS
- 99.97% of bot checkout attempts blocked across 3 simultaneous drop events
- 4.2× increase in successful purchases by real customers vs previous drop
- $0 in chargebacks from bot-originated orders (previously averaging $180K/drop)
- Scalper infrastructure identified across 23 unique ASNs and 6 residential proxy networks
- Session fingerprint library updated within 40 minutes of detecting new evasion variant
AI Data Platform Eliminates Unauthorized Data Scraping
An enterprise market intelligence vendor faced continuous scraping by LLM web agents and competitor bots attempting to harvest proprietary dataset APIs. Axiom Zero Layer 05 (Biometric Motion Kinematics) and JA4+ TLS signature auditing identified automated scraper requests before database queries executed, reducing backend database load by 48%.
RESULTS
- 12.4TB of proprietary training data protected over 90-day evaluation period
- 99.94% of AI scraping agents blocked without impacting legitimate API integrations
- 3 distinct LLM agent architectures identified and catalogued in the threat intelligence vault
- Infrastructure cost reduction: 62% lower than previous WAF-only approach
- Zero customer support tickets related to false positives during evaluation period