Security becomes a shared job as AI widens the attack surface

That shift has moved the relationship from infrastructure plumbing to joint defense of models, agents and the services ecosystem around them. Customers building applications on cloud infrastructure are also building agents there, widening the ground both companies must protect, said Mona Chadha, director of AI and strategic partnerships at AWS.

Chadha and Amanda Adams, senior vice president of global alliances at CrowdStrike, spoke with theCUBE at Fal.Con. Chadha said AWS and CrowdStrike continued to build innovations together over the decade. "Over the decade, what we continued to do was build innovations together," she said. "We’ve evolved from machine learning to generative AI to now agentic and building these agents. There’s a lot of opportunity there, but there [are] also a lot of threats."

Neither company is trying to solve AI security alone. CrowdStrike’s answer is Project QuiltWorks, a coalition that includes cloud providers, systems integrators, technology suppliers, law firms and insurance partners. Adams said the newly announced Falcon IQ tool, built on AWS, lets partners drive an assessment, highlight priorities and use AI to reduce the time from discovery to remediation to minutes.

CrowdStrike counts more than 900 technology integrations across its partner ecosystem, Adams said, and channel and services partners are creating new revenue streams on top of the joint AWS-CrowdStrike stack, according to Chadha.

"Build everything with security in mind. Security is job zero," Chadha said. "As long as we have that security layer built in, that’s where we can ensure that we have the proper guardrails, that then ensure that these agents don’t go out of that sandbox." She added that underlying security discipline has not changed as agent-level controls arrive.