Quarry Systems Donghan Yoon Exclusive Director

"Security Operations Centers are increasingly burdened by high alert volumes, repetitive event analysis, false-positive validation, report writing, and response decision-making.
Generative AI and AI agents can automate and enhance SOC tasks such as log summarization, alert triage, threat intelligence analysis, and report generation.
However, granting AI excessive authority can introduce risks such as prompt injection, sensitive data exposure, incorrect decisions, and uncontrolled automated responses.
This makes harness engineering essential to control what data AI can access, which tools it can use, and how far it is allowed to act.
This session presents practical approaches to designing controllable AI security operations agents, including use cases, safety controls, and performance measurement methods."