BALANCECHEAT / AI Scenario Analysis
Compare coherent cases, one model at a time.
Use your own AI to build and inspect scenarios of the exposed financial model. Keep Base, Upside and Downside assumptions distinct, with pending AI changes visible in each case.
Your model · Your AI · You stay in control
Loading the statement preview…
Open the full financial model →Continue with your AI in the full model
This is a focused preview with sample data. Open the full financial model, choose AI, and connect your own provider to analyze or edit your model. Your saved models are unchanged by this preview.
A useful starting prompt
Try this with your provider in the full model:
Compare Base and Downside. Identify the assumptions that explain their cash difference before making any edits.
Make each case tell a consistent story
A scenario should express an operating narrative rather than a collection of unrelated percentage changes. Ask your AI to inspect the available scenarios, then align revenue, hiring, working capital and investment with a stated downside or upside assumption.
The AI can address every scenario belonging to the exposed active model. Other saved models remain outside that connection. Switching the exposed model invalidates the old context and revokes an external connection.
Review across scenarios without losing the source
Pending AI changes aggregate into their scenario and the Scenarios control. Review all AI changes groups them by scenario and model area. Review Mode moves between direct edits, including edits in a scenario that was not previously visible.
Accept one assumption, accept a group, or inspect all changes before accepting them. Manual editing takes precedence over a pending AI edit to the same field.
Bring Your Own AI
Connect a compatible AI client to work with a structured financial model. ChatGPT, Claude and Gemini offer external connections where their product, region and plan allow them. Your own API key remains an advanced option for chat inside BalanceCheat. BalanceCheat provides the calculation engine; your connected AI provides the intelligence.
Your financial model stays stored in your browser. When you use AI, your prompt and the requested model context are sent to your provider. The relay does not retain provider keys. External connections temporarily hold tool requests and results while your browser executes them.