Defeating Advanced Fraud with Project ERICA

The Challenge

A major platform was facing a sophisticated wave of synthetic fraud and credential stuffing. Attackers were using distributed proxy networks to rotate IPs, perfectly bypassing standard WAF rate-limiting and managed rules. To counter this, the engineering team had implemented aggressive CAPTCHAs, which successfully slowed the bots but severely damaged conversion rates and frustrated legitimate users.

The Blueprint

Aspida Labs deployed Project ERICA as an intelligent, post-WAF overlay natively on the edge to close the gap between application security and business logic fraud.

• Zero-Risk Deployment: ERICA was initially deployed in "Shadow Mode". For 7 days, it silently scored payloads and business logic inline at the edge in log-only mode, establishing a behavioral baseline — without blocking a single request.
• Precision Risk Scoring: We activated The 'BRAIN' — ERICA's Bayesian risk model, applied inline at the edge. Instead of relying on static IP reputation, it correlated identity and behavioral impossibilities to flag fraud rings.
• Cryptographic Federation: We secured the authentication endpoints, ensuring that only trusted payloads could interact with the origin database.

The Outcome

By moving the anti-fraud intelligence directly to the edge, the platform achieved massive security gains while actually improving the user experience.