Standard Bank Group worked with AWS Professional Services to build an automated assessment that checks each generative‑AI use case against international standards and the bank’s own policies. The pipeline runs in the bank’s AWS account, uses Amazon Bedrock as an “LLM‑as‑a‑judge”, and outputs reports and a dashboard for the risk committee. The bank says the system reduced a typical manual review from two weeks to roughly two hours and improved risk scores by 14% after remediation.
Why it matters
Compliance teams can approve AI projects faster while keeping evidence‑based oversight.