Production Spam Detection Across Multiple Sites

We built an AI-powered spam detection layer that screens form submissions across multiple websites operated by an Irish IT services firm. The system classifies incoming submissions as legitimate or spam using GPT-based intent analysis, replacing brittle keyword-based filtering that produced too many false positives and missed sophisticated spam.

The system has been running continuously in production for over 12 months, processing form submissions from multiple end-client sites under a single integration layer.