COMPEL Glossary / sensitivity-analysis
Sensitivity Analysis
Sensitivity analysis is a technique that tests how changes in key assumptions affect the outcomes of a business case, financial model, or risk assessment.
What this means in practice
For AI investments, sensitivity analysis identifies which variables most significantly impact projected returns: What happens if adoption is 30% lower than projected? What if data quality improvements take twice as long as planned? What if compute costs increase by 50%? By modeling a range of scenarios for each critical variable, sensitivity analysis helps transformation leaders understand the robustness of their business cases and identify the assumptions that carry the greatest risk. In the COMPEL Model stage, ROI projections must be stress-tested against at least two alternative scenarios, and the stage gate criteria require that no single use case accounts for more than 40% of projected cycle value.
Why it matters
Sensitivity analysis reveals which assumptions carry the greatest risk in AI business cases by testing how changes in key variables affect projected outcomes. Without it, organizations commit to investments based on single-point estimates that may be highly sensitive to adoption rates, data quality timelines, or compute costs. Understanding sensitivity enables better risk management and more robust investment decisions.
How COMPEL uses it
During the Model stage, ROI projections must be stress-tested against at least two alternative scenarios, with stage gate criteria requiring that no single use case accounts for more than 40% of projected cycle value. The Evaluate stage compares actual outcomes against sensitivity scenarios to calibrate future projections. The Learn stage uses sensitivity data to improve the accuracy of business case modeling in subsequent cycles.
Related Terms
Other glossary terms mentioned in this entry's definition and context.