Silent Agents Are Already on Your Site
Akamai Technologies, Monday, October 26th, 2026
Akamai argues websites should govern unverified AI agents by purpose and endpoint rather than blocking or allowing them outright.
The Akamai blog post says AI agents acting through ordinary browser or customer sessions, without verifiable identity, are already visiting websites, and that blanket blocking or allowing both fail.
It notes that major banks have restricted external agents from customer accounts and that retailers lose sales when anti-bot controls block shopping agents.
The author proposes selective action, letting agents browse catalogs while applying stricter scrutiny to checkout, login, and account changes.
Akamai's approach separates training crawlers, search crawlers, and agents, uses behavioral risk scoring for unverified sessions, and points to verification through standards such as Web Bot Auth as the route to wider access.