Akamai uses machine learning in its Account Protector product to assess in real time whether a login or account action is coming from the genuine account owner or an impersonator, generating a risk score that customer businesses can use to block or challenge suspicious requests. The product launched in June 2021 and received expanded capabilities in October 2024.

Account Protector builds an individual behavioral profile for each user over time, capturing typical devices, locations, networks, and activity patterns. For each new login or account action, the ML model compares the request against the stored profile and cross-references network-wide reputation signals from all Akamai customers to calculate a risk score. In October 2024, Akamai added lifecycle protection covering post-login activities including account updates, password changes, and payment transactions.