New AWS Bedrock Dimensions: From Buried ARNs to Actionable Cost Attribution

AWS Bedrock inference profiles provide clean organizational structure for AI usage, but in the Cost and Usage Report (CUR), all that structure collapses into opaque ARN strings that are nearly impossible to filter or group. The core challenge is distinguishing spend that flows through inference profiles from untracked spend, and then breaking down costs by engineer, team, or environment. A real-world example shows how assigning per-engineer inference profiles via Terraform enabled a FinOps team to identify power users and cut costs 40% by switching them to flat-rate licenses. Finout has released two new dimensions — Inference Profile ID and Inference Profile Type — that automatically parse Bedrock ARNs to enable actionable cost attribution without custom scripts.