How to model financial scenarios and sensitivities with AI

Financial scenario modeling is the practice of building multiple versions of your future — a base case, an upside, a downside — and stress-testing your assumptions before a decision locks you in. For most operators, it shows up before a hiring push, a pricing change, a fundraise, or a vendor contract. The question is always some version of: what does our cash position look like in 12 months if X happens instead of Y?

AI feels like a natural fit here because the core mechanics are logical, not creative. You're essentially asking: given these inputs, what do the outputs look like under different assumptions? Large language models are decent at arithmetic, can hold multiple scenarios in context at once, and can explain the logic behind a number in plain English — which is often what you need to communicate findings to a board or co-founder, not just produce a spreadsheet.

ChatGPT, Claude, and Gemini can all contribute meaningfully to this workflow. They can build scenario tables from numbers you paste in, apply sensitivity rules, flag which assumptions move the needle most, and format results in a way that's readable to non-finance stakeholders. The ceiling is that they work on data you bring to them, not data they can pull themselves — and they don't remember anything from the last time you ran the exercise.