AI for security and AIOps
Using AI to do security work — triage, investigation and remediation.
This hub is for a different use of AI than the rest of the site — not securing AI workloads, but using AI to do security work itself: triage, investigation, and remediation assisted by a model rather than done by hand end to end. It's a natural next step from lessons already published elsewhere: automated remediation with EventBridge and Lambda and turning findings into APRA narrative with Bedrock both already lean on AI for part of the loop, in the detection-response and compliance-evidence hubs respectively.
There's no lesson filed under this hub yet. It exists first so the first AIOps-focused post has somewhere to land, rather than being squeezed into an adjacent hub where it doesn't quite fit. Triaging 500 Prowler findings into a plan is, right now, a manual process — reading a scan, weighing severity, deciding what to fix first. That's the kind of work this hub is for: what changes, and what doesn't, when part of that judgement is delegated to a model. Expect this space to fill in as that work is published.