Cloudflare has introduced Adaptive Intelligence, a continuous detection engine built into its Bot Management product, which the company says is designed to raise the cost and effort involved in running automated cyber attacks.
The engine draws on data from Cloudflare’s global network and learns from what the company describes as the meta-signals of live traffic. It then produces short-lived rules aimed at specific threats.
Go deeper with GlobalData
Access deeper industry intelligence
Experience unmatched clarity with a single platform that combines unique data, AI, and human expertise.
According to Cloudflare, the system analyses more than a trillion web visits each day and reacts when attackers get past existing defences.
Cloudflare chief technology officer (CTO) Dane Knecht said: “Building taller walls fails when the cost of scaling an attack is effectively zero. To stop modern bot threats, you have to flip the economics on the attackers. Traditional defences offer a static target that threat actors can systematically solve.
“Cloudflare’s Adaptive Intelligence creates a moving target—rendering an attacker’s engineering efforts obsolete before their operation can ever achieve scale.”
Cloudflare argues that launching automated attacks once demanded advanced engineering skills, but that AI and inexpensive online tools have lowered the barrier.
The company said attackers with modest budgets can rent networks of compromised devices, hide behind residential internet addresses and use free software that imitates simple human behaviour. Defenders, by contrast, must block every attempt without turning away genuine visitors.
The company also points to a timing gap. Many security tools depend on scheduled updates that can take weeks or months to reach customers, while attackers may change tactics within hours or minutes.
Cloudflare lists several functions for the new engine. Its machine learning model retrains on live traffic, adding new bot frameworks and bypass methods without waiting for manual releases.
Rules rotate frequently, which the company says stops attackers mapping defences or building on earlier progress.
The engine examines behaviour across several timeframes at once, with the aim of detecting slow credential stuffing and scraping campaigns.
Adaptive Intelligence also combines browser-level session signals from Cloudflare Precursor with global edge telemetry, moving beyond binary tests.
Updates are tested against live traffic before deployment, which Cloudflare says checks accuracy and limits false positives without downtime.
Last month, Cloudflare launched Radar Researcher in beta, an AI tool that lets users query internet traffic and trend data in plain language.
