Ingrid Larsen checks her forecast dashboard each Wednesday morning before the executive team meeting. It took her 16 months and two failed vendor relationships to build something that didn't require weekly manual corrections to stay aligned with actual business performance.

The Automation That Made Things Worse

Her first dashboard automatically extrapolated trends from historical data, which worked until market conditions shifted. The second version incorporated more variables but became a black box that nobody trusted when it projected numbers that contradicted what division leaders were seeing. Ingrid kept trying to make algorithms smarter when the real problem was that automation can't replace business judgment.

Separating Data Work From Decision Work

She redesigned the system to automate data gathering and standardization while keeping forecast assumptions manual and visible. Division leaders now input their projections with required explanations for significant changes. The dashboard aggregates and checks for inconsistencies but doesn't override human judgment. Her forecasts now stay within 6 percent of actuals because the system supports decision-making instead of trying to replace it. The automation finally works because it knows what not to automate.