What problem does it solve? Founders often build a forecast and then never close the loop on it, so plans drift from reality and the same wrong assumptions repeat month after month. This Skill runs the monthly learning loop: it compares actuals against forecast, prior month, and prior year, explains each variance in plain language, and turns the findings into owned, dated next steps. ## Core Features & Use Cases - Data-Completeness Gate: Blocks the review when forecast or actuals are missing for any headline metric, so variance analysis is never presented on half-populated data. - Three-Baseline Variance Analysis: Compares Revenue, Expenses, Net Profit, and Cash against forecast, previous month, and previous year, classifying each as higher, lower, or about the same using a configurable tolerance band. - Top Movers Diagnosis & Co-Movement Checks: Ranks line items driving each variance and detects when a favorable cost variance is actually an artifact of stalled activity rather than savings. - Use Case: A founder says "Close out July — how did we do vs plan?" The Skill verifies the books are closed, computes variances, flags that zero expenses alongside zero revenue means inactivity (not efficiency), checks risk thresholds and assumption-miss streaks, and produces dated next steps with owners. ## Quick Start Ask the assistant to run my monthly business review for last month comparing actuals to the forecast and telling me what to change.
