Applying FinOps to GenAI in a Mature Tech Organization

When Intuit QuickBooks pivoted to Agentic AI for “Done for You” customer experiences, our financial landscape shifted overnight—from predictable cloud infrastructure to complex LLM unit economics, semantic databases, and expensive telemetry.

This session reveals the FinOps journey behind Intuit’s Accounting Agent. We prioritized model accuracy and time-to-market while building a cost optimization framework that scaled with each development phase. Engineering and Finance partnered to identify strategies that controlled spend without sacrificing velocity or customer value.

Key takeaways:

• The LLM Cost Matrix: Evaluating commercial models versus fine-tuned internal models, and how prompt context tuning impacts the bottom line
• Invisible costs: Managing standard cloud services plus unexpected expenses from logging and data freshness for analytics
• Architectural FinOps: Optimizing semantic databases for AI workloads, including capacity planning and bottleneck management
• Forecasting the unknown: Managing stakeholder expectations when traditional forecasting models fail

You’ll leave with a blueprint for applying the FinOps Inform-Optimize-Operate framework to GenAI—moving beyond cost-cutting to ROI-driven engineering.