Connectivity cloud company Cloudflare has launched Adaptive Intelligence, a machine learning tool that detects malicious bots and continually adapts to keep pace with new attack methods.

“Adaptive Intelligence is designed to operate autonomously, learning from live traffic and signals across Cloudflare’s network to identify the likelihood of automated abuse,” Bryan Becker, director of product, application security integrity and trust solutions at Cloudflare, told FinAi News.

“It continuously runs through a loop of observing, training, deploying and validating so defenses can adapt as attacks evolve without requiring teams to manually push updates.”

Adaptive Intelligence, launched Aug. 31, makes pursuing an attack too slow and costly for attackers by replacing a deterministic bot detection approach with a fluid, continually evolving strategy, according to Cloudflare.

“In this age of AI, where exploitation is fully automatic and continuous, you cannot patch your way out of this mess,” Mahdi Abdulrazak, chief executive and co-founder of cybersecurity platform Dawnguard, told FinAi News. “Real-time, adaptable defenses are critical.”

Determinism’s obsolescence

A deterministic program is, by definition, not adaptable. It will always give the same output to the same input.

“The challenge with rule-based systems is that they hand the attacker a stationary target,” Chris Pope, senior product manager for Adaptive Intelligence, wrote in an Aug. 31 blog.

“Bot detection has always answered a new attack technique by writing a rule to catch it,” he said. “That works — until the attacker studies the signal, learns how to circumvent it and forces another rule to be written.”

Because deterministic systems are static, attackers learn how to beat them by throwing multiple attacks at a system and studying its response, Pope said.

Moving target

Adaptive Intelligence undercuts this by weighing multiple signals and not visibly reacting to flagged signals.

“Essentially, we make it harder for an attacker to reverse-engineer how and why they were being detected,” Becker said.

Adaptive Intelligence has three components:

1. Improving itself: The system regularly updates itself rather than waiting for a scheduled system update.
2. Disposable rule generation: The system not only adds rules as they become necessary but discards them when they become irrelevant. However, the rule is never truly forgotten; the system can recall it later should an attacker return to a previous form of attack.
3. Learning from real user traffic: When the system learns it has given a false positive, that correction becomes a training signal.

The first component launched with the product, and the others will be released incrementally, Becker said.

How it works

Adaptive Intelligence observes, trains, deploys and validates in a loop:

• Observe: The system aggregates signals from across its network.
• Train: The system studies the data and refreshes its training set often. The refresh happens as needed, not at a set time, Becker said.
• Deploy: New models roll out on their own. Customers do not need to upgrade their systems.
• Validate: Before a new version becomes the primary defense, it runs in shadow mode alongside the current one. If the new version is not an improvement, it will not go live.

Human analysts monitor and enhance the model, alongside automation, and are alerted if something goes wrong, Becker said.

FinAi Lending Summit, set for Oct. 7-8 in Las Vegas, will include speakers from Fifth Third and Capital One as well as a fireside chat with U.S. Bank Senior Vice President – Lending Operations Group Manager Suzanne Rathbun. To learn more about the 2026 event and register, visit here.