Case Study: a large IT company improves AI transparency and risk control with Code4thought’s network monitoring audit

Large IT Company - Customer Case Study

a large IT company engaged Code4thought to conduct an AI audit for its Network Monitoring Model, which was a complex, unsupervised system used to detect internal security threats. The challenge was to evaluate and explain the model's behavior, as it operated without human-labeled data and its self-identified threats lacked a clear ground-truth for validation.

Code4thought used its proprietary technology to address this. The solution included an Accountability Evaluation checklist, a Bias Testing mechanism, and the M.A.SHAP explanation algorithm. This process identified key areas for improvement and established new governance processes. As a result, the client operationalized the explanation tool for transparency, shifted the model to a supervised learning approach, and implemented a new risk management framework, ultimately increasing confidence in the product's robustness and the maturity of their organization.