The Finance Leader Who Audited AI Spend — And Won

Emily Allen, Director of Finance and Accounting at CloudZero, brings a perspective that is uncommon in FinOps discussions: she is the finance practitioner who actually runs the analysis herself. In this episode, she walks through how she used CloudZero to audit internal AI usage, identified over-reliance on high-cost models, and initiated conversations that led to $30K in annualized savings — all without triggering the defensive reactions that finance-driven cost reviews often produce. For finance and FinOps leaders trying to get ahead of AI spend before it becomes a budget problem, this episode is a concrete starting point.

This episode covers how Emily Allen used CloudZero data to analyze AI tool usage across her organization, identify model-level inefficiencies — specifically the over-reliance on Claude Opus for tasks better suited to Sonnet — and generate approximately $30K in annualized savings through targeted conversations with internal teams. She also addresses the broader challenge finance leaders face with AI spend: the shift from centralized procurement to distributed, card-swipe adoption, and what that means for visibility and accountability.

Key Takeaways

Finance teams do not need to wait for a formal FinOps program to start finding AI spend savings.
Visibility into usage data at the model and user level, combined with direct peer conversations, can produce measurable results quickly.
AI spend governance requires a different posture than traditional software procurement. Because adoption is decentralized and frictionless, the finance function needs to shift toward continuous education and cross-functional relationship-building rather than periodic approvals.

About the guest

Emily Allen, Director of Finance and Accounting at CloudZero, brings a perspective that is uncommon in FinOps discussions: she is the finance practitioner who actually runs the analysis herself. In this episode, she walks through how she used CloudZero to audit internal AI usage, identified over-reliance on high-cost models, and initiated conversations that led to $30K in annualized savings — all without triggering the defensive reactions that finance-driven cost reviews often produce. For finance and FinOps leaders trying to get ahead of AI spend before it becomes a budget problem, this episode is a concrete starting point.