How Cisco defends at machine speed with OpenAI

Attackers are using AI to automate attacks and build exploits faster. The gap between when a flaw can be exploited and when it gets patched is where breaches happen. Cisco says pairing its network telemetry with OpenAI's models helps close that gap.

Cisco President and Chief Product Officer Jeetu Patel describes security pushed down to every microservice, container and cluster, with AI correlating signals that look harmless alone. He credits OpenAI's Daybreak and GPT-6 Astra with helping Cisco find and patch vulnerabilities faster, and says the tools slot into existing scanning workflows.

The two companies are also working through the Daybreak Defense Network on faster exposure triage and control enforcement in Cisco Cloud Control.

We agree with the premise. If attackers get speed and scale from AI, defenders who refuse it are choosing to lose. Patel's line that using AI for defense is a "prerequisite for staying safe" is hard to argue with, and correlating weak signals across a network is exactly what these models should be good at.

The agentic piece is the most forward-looking part. Patel separates a malicious attack from an agent pursuing a legitimate goal in an unintended way, and argues teams need security telemetry plus observability to tell them apart. That's the right problem to be naming now, before autonomous agents are everywhere.

Still, this is a customer story published by the vendor. There are no metrics: no reduction in time to patch, no count of vulnerabilities found, no false-positive rate. Calling the gains a step-function improvement is a testimonial, not a measurement. The privacy claims, zero data retention and verifiable isolation, are asserted rather than shown.

The risk if this scales is concentration. When the network vendor and the model vendor become the default defensive brain for a large share of enterprises, a flaw or outage in that combination becomes a shared failure. Defense at machine speed also means mistakes at machine speed, like an automated control that isolates the wrong service. We want to see how humans stay in the loop when the response is that fast.

Daybreak and GPT-6 Astra are OpenAI security initiatives that Cisco says it uses to find and remediate vulnerabilities. The Daybreak Defense Network is the joint effort Cisco and OpenAI describe for helping security teams triage exposure and enforce controls through Cisco Cloud Control.