CargoSprint customers regularly requested detailed spend breakdowns — invoice-level data, year-over-year comparisons, facility performance, and transportation mode splits. Each report was built manually in Excel, one customer at a time. At scale, this took approximately one week of analyst time per reporting cycle.

The problem wasn't capability — the reporting structure was well-defined. The problem was that the same logic was being re-executed by hand, over and over, for each customer on each cadence. The fix was to stop doing that.

• Built a Python pipeline that generates 40+ unique recurring customer spend reports automatically on a scheduled basis
• Report content: invoice-level detail, YTD and rolling-13 comparisons, peak volume periods, facility-level breakdowns, transportation mode splits, and fee category analysis
• Delivery format: Excel workbooks — matching what customers in a blue-collar freight industry actually use and prefer
• Built a Python UI for onboarding new customers — adding a customer triggers automatic report generation on their chosen cadence (weekly, bi-weekly, or monthly)
• No analyst action required per cycle after setup — the system scales as customers enroll

The pipeline replaced approximately one week of manual work per reporting cycle with fully automated generation — a 90% efficiency gain. New customers self-enroll via the onboarding UI with no analyst intervention per cycle.

The reporting program expanded over time as more customers enrolled without adding work to the analyst's plate. The system scales; the manual process did not.