Cloudflare launches Adaptive Intelligence for bot attacks

Cloudflare has launched Adaptive Intelligence for its Bot Management service, designed to respond continuously to automated cyber attacks.

The new detection engine is built into Cloudflare's existing bot mitigation product and draws on traffic patterns across its network. It learns from live internet traffic and generates short-lived rules intended to make automated attacks harder to sustain.

Businesses have faced a rise in automated abuse as artificial intelligence tools and low-cost online services lower the barrier to launching attacks. Security vendors have increasingly focused on bot traffic that imitates legitimate users, spreads requests across many devices and changes tactics quickly enough to evade products that rely on periodic updates.

Cloudflare said Adaptive Intelligence analyses more than a trillion web visits a day. That scale allows it to retrain its machine learning models continuously rather than waiting for manual intervention or scheduled releases.

The product is aimed at attack types that can be difficult to detect with traditional defences, including credential stuffing, scraping campaigns and what security teams often call low-and-slow attacks, where malicious activity is spread over longer periods to avoid triggering standard thresholds.

How it works

According to Cloudflare, the system combines browser-level session behaviour signals from its Precursor technology with telemetry from its wider network. This approach is intended to help distinguish legitimate users from malicious automation without relying on simple binary checks.

Another part of the design is the use of temporary detection rules tailored to particular threats. These rules rotate frequently, aiming to prevent attackers from studying a static defence and refining techniques against it over time.

Updates are also tested against live traffic before deployment. The process is intended to reduce the risk of false positives, a persistent concern for online businesses that need to stop abusive traffic without disrupting genuine customers.

Economic shift

Cloudflare framed the launch as a response to the changing economics of cyber crime. As attack tools become cheaper and easier to access, defenders face pressure to identify and stop large volumes of automated traffic, while attackers can keep trying until they find a gap.

That imbalance has become more visible in sectors with heavy consumer traffic, such as retail, travel, financial services and media, where bots are used for account takeover attempts, content scraping, inventory hoarding and other forms of abuse. In those settings, the cost of blocking an attack incorrectly can be almost as damaging as missing one because legitimate customers may be turned away.

Cloudflare argued that older approaches can struggle because they depend on updates that may take weeks or months to reach customers. By contrast, Adaptive Intelligence is intended to adapt as new frameworks and bypass methods emerge in live traffic.

Dane Knecht, Chief Technology Officer, Cloudflare, outlined the company's view of the shift in attack patterns.

"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 defenses 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," said Dane Knecht, Chief Technology Officer, Cloudflare.

The launch adds to a broader industry push to use live traffic analysis and machine learning in cyber defence. Vendors across network security, fraud prevention and identity protection have been updating products to cope with attacks that blend automation with behaviour designed to look human.

For customers already using Cloudflare Bot Management, the significance of the new feature will depend on whether it can improve detection without increasing friction for legitimate visitors. In bot defence, accuracy matters as much as speed, particularly for organisations that handle large volumes of logins, searches and transactions every day.

Cloudflare said Adaptive Intelligence functions as a single learning system that reacts to threats as they happen, rather than as a set of fixed signatures that attackers can probe repeatedly.