If a cyber breach can happen in 25 minutes but it takes days to contain, are organisations already too late?

New cloud security findings from Palo Alto Networks point to a stark reality: attackers are now operating at machine speed, fuelled by AI. Nearly every organisation surveyed reported at least one attack on its AI systems in the past year. APIs are becoming key entry points. Identity weaknesses are driving incidents. And security teams are still juggling an average of 17 tools while trying to respond in real time.

As companies race to deploy faster using AI and automated development tools, they may be expanding their exposure just as quickly. Prevention remains critical. But resilience and response speed are fast becoming the true differentiators.

Elad Koren, Vice President of Product Management for Cortex Cloud at Palo Alto Networks, joins us to examine what breaks first under machine-speed attacks and what needs to change to keep up.

We explore:

• Why the shift from prevention to response speed is becoming the new baseline.

• What AI poisoning really means, and how intentional sabotage of AI systems could play out.

• Why APIs have become both the connective tissue and a growing vulnerability in modern cloud environments.

• The identity challenge in cloud and AI systems, and why identity now functions as the last line of defence.

• Why manual data reviews are becoming a liability at cloud scale.

• Whether the traditional separation between cloud security and the SOC still makes sense in the AI era.

This is not just about stronger tools. It is about whether organisational design, security workflows, and accountability structures are built for threats that now move at machine speed.