Darryl Strucko

Problem

Tier 1 analysts are overwhelmed by alerts and often lack confidence to escalate correctly. AI can act as a trusted advisor, making recommendations the analyst can question directly. This project validates the AI interaction model with real users and provides the ability to test and calibrate the AI's accuracy.

Approach

•Seven case scenarios test AI reasoning across different attack types.

•Claude gives evidence-grounded, explainable answers to analyst questions.

•Test Rig and Calibration Rig evaluate and refine AI accuracy.