@Alix Lefort: This is an interesting question. The approach to scenario planning can vary significantly based on the industry, as well as the primary regions driving revenue.
While global events create broad economic volatility, their impact may not be uniform worldwide. Some industries and geographies tend to be more resilient, while others are far more exposed to sudden and unpredictable shifts.
Now that I put in a disclaimer, here are 3 examples, of a more practical approach to scenario planning in today's environment - where the picture is incomplete and conditions keep moving :
1/ Keep it simple: I agree with your point about simplifying. What's worked well - especially in uncertain times like these - is narrowing it down to a small number of meaningful scenarios, aligned to the overall financial strategy and company objectives.
Within each scenario, cut through the noise, focusing 3- 5 variables that truly move the business.
2/ A reasonable buffer: Regular touchpoints and relationships with business/regional leaders provide a much clearer picture of ground realities and market sentiment. This qualitative insight in turn, gives a practical buffer within scenarios planned.
In volatile environments, that buffer often makes the difference between a scenario that looks good on paper Vs one that holds up in practice.
3/ Use AI with professional skepticism: If you already have AI as part of the FP&A toolkit, it can be a strong enabler - especially for dynamic driver based modeling and sensitivity analysis (e.g., Revenue, volume, FX movements)
That said, it still requires professional oversight. AI outputs are often based on historical trends/seasonality/ time lag between getting updated results based on unforeseen circumstances), which may not fully capture sudden shifts.
Interpreting those outputs, and aligning them with what matters strategically, is where finance judgment remains critical.
Before diving too deep into any scenario, I've found it useful to pause and ask:
"What decision will this scenario actually inform" .
If that is not clear, the model may need simplifying further
Hope this helps and great to see this discussion getting the attention it deserves 😊
------------------------------
Joseph Azavedo ,ACMA, CGMA
Wroclaw
------------------------------