Akamai uses an AI framework in its Bot Manager product to automatically distinguish human web visitors from malicious automated bots across customer websites, apps, and APIs, assigning each request a bot-likelihood score in real time. This capability has been in production for multiple years and handles an average of 40 billion bot requests per day.

Bot Manager's AI framework operates inline on Akamai Connected Cloud, analyzing requests at the edge before they reach customer origin servers. It uses machine learning models, behavioral analysis, device fingerprinting, and browser anomaly detection to assign a score from 0 (human) to 100 (bot). The system includes per-customer deep learning models that learn attack patterns specific to individual high-profile customers and auto-tunes over time. In October 2025, Akamai announced advanced bot detection enhancements including global recognition of trusted users across its entire customer network.