Writing
Writing on agentic AI, autonomous security operations, software delivery, compliance engineering and the practice of security leadership. Everything here comes from systems we run live - not from slideware.
We ran a fully autonomous SIEM-driven SOC in production for six months. Agents closed 95% of the workload end to end with zero false positives, and the analyst role stopped existing. Here is the architecture that made it safe.
One layer in front of every AI request, choosing which model answers it. It kept 700+ employees working through outages at two providers and cut AI running costs by up to 80%. The architecture, the trade-offs and the lessons.
Security posture scores move when you make detection and automation do the work. A quarter-by-quarter account of the fastest posture transformation we have run - without headcount growth.
ISO 42001 and the EU AI Act don't tell you how to govern autonomous agents in production. After shipping a fleet of them, here is the framework we would enforce as a CISO - guardrails, agent identity and human-in-the-loop by design.
Work with FeatherON
Book a focused hour on the problem you are actually facing - agentic AI risk, a compliance wall, a build that has stalled, or a board that needs answers. You will leave with a plan, not a pitch.