A good simulation lets you learn by changing something. You make a choice, see what happens and try again. After a while, you develop intuition for the system.
Financial models can teach you in the same way. Move a hiring date or change a payment assumption and watch the result. The easier that experiment is to run, the more questions you can explore before making a real commitment.
That is the part of playing a game worth bringing into business software: a close connection between an action and its consequences.
Make the feedback useful
A financial report gives you a result. To develop intuition, you need to see how the result responds to a change.
Take a simple cash example. A company starts April with $250,000. Monthly spending is $100,000, regular customer payments are $60,000, and a separate $180,000 invoice payment is expected in April. Those are the only cash flows in this example.
Move that invoice payment to June:
| Month-end cash | Payment in April | Payment in June |
|---|---|---|
| April | $390,000 | $210,000 |
| May | $350,000 | $170,000 |
| June | $310,000 | $310,000 |
June ends in the same place. The cash available during April and May is very different.
That distinction is easy to miss when you only see the final balance. A view that lets you move the payment date and compare the paths makes it much easier to grasp. Lower the opening cash and you can explore when a delay starts to create a funding problem.
The monthly view still has a limit: it doesn't tell you whether payroll comes before or after a payment within the month. If the decision depends on that order, use the actual dates.
Build the controls around the question
The most useful control in that example is the payment date. It corresponds to something the person running the business understands and can investigate.
Other questions need different controls. A hiring decision needs a start date and a view of the work the hire is expected to produce. A pricing decision needs assumptions about how customers respond, with enough detail to distinguish existing contracts from new sales.
Putting every possible input on the screen makes the relationship harder to see. Start with the decision, then expose the assumptions that could change it.
The reader should be able to point to the change they made and explain why the result moved. If they can't, adding another chart is unlikely to help.
Your company can have its own interface
CFO.ai's Canvas blocks let Ari build custom visual experiences for particular questions. They can present model data alongside controls and explanations, so the way you explore the decision can fit the decision itself.
For the payment example, that could mean a payment on a timeline with the cash balance underneath. For a hiring plan, it could mean showing when each role starts and how the modeled cost develops over time.
These are interfaces you can return to and work with. As the business changes, the questions and assumptions can change with it.
That is a useful reason to generate software around a model. A generic report has to serve many readers and many decisions. A custom Canvas can focus on the question in front of you.
Trying a control can change a local simulation without changing the saved company plan. Exploration and saving should be clear to the person using it, especially when several people depend on the same model.
The rules of a business are partly unknown
A game's designer decides how the world responds. In a business, some of the most important relationships are uncertain.
You can calculate the payroll cost of an agreed start date. Estimating how much revenue the hire will produce takes evidence and judgment. A model can show the consequences of that estimate; it cannot make the estimate true.
That uncertainty gives experimentation a purpose. Try a slower ramp or lower sales output. See whether the decision still makes sense and which assumption changes the answer first.
Ari can help build and examine the model. Your knowledge of customers and the work itself remains part of the process. When an assumption is weak, the useful next step may be to gather evidence before committing.
Leave with a better feel for the business
After exploring the payment example, you understand why the same June balance can hide very different conditions in May. That knowledge travels with you into a conversation with the customer or a decision about spending.
This is what makes interactive business software worth building. It gives you repeated chances to connect a choice to a result while the choice is still reversible.
Over time, you develop intuition for your own company. You can explain what matters, test a different plan and recognize when the facts have changed enough to require another look.
CFO.ai editorial adaptation of Siqi Chen's original Runway essay. Updated September 18, 2026.