Finance teams often drown in the repetitive grunt work of closing the books, a reality Snowflake’s FP&A department faced head-on. Hours were lost each month assembling spreadsheets, aligning data, and chasing budget versus actual analysis before any meaningful strategic work could begin. This manual grind, a common pain point for FP&A professionals hired to explain business performance, was ripe for disruption. The traditional process meant delays, visibility gaps, and a constant battle against formatting issues and organizational changes. Actuals from systems like Workday often had a two-hour refresh lag, further bottlenecking timely insights. This was the exact problem Snowflake’s internal finance team set out to solve using Snowflake Cortex Code, or CoCo.
From Static Workbook to Live Dashboard
The appeal of CoCo lay in its native integration within Snowflake. This meant the AI coding agent already had context on the company’s data models, mappings, and forecast tables. Instead of onboarding data into a separate environment, analysts could describe their desired variance view in plain English, with CoCo translating it into action against existing Snowflake objects. This drastically shortened the path from concept to a functional tool. A clean hierarchy was established in Snowflake, enabling full drill-down capabilities from consolidated P&Ls to individual cost centers. CoCo then accelerated the build of a budget vs. actual dashboard in Streamlit, entirely replacing the static Excel workbook. The pipeline now runs continuously, pulling actuals directly into Snowflake without manual intervention, eliminating the lag and enabling real-time variance review during close.
