A spreadsheet with a fixed 7% produces one number, and that number is almost certainly wrong: no market delivers 7% every year. A Monte Carlo simulation draws thousands of plausible return sequences and counts how many of them end well.
The output is not "you will have $480,000" but something far more useful: "in 85% of scenarios the money lasted thirty years; in 15% it ran out early". That turns a forecast into a probability, which is the right way to think about a financial future.
Its obvious limit: the simulation only knows what you told it. If the returns and volatilities you feed in are optimistic, you get a thousand optimistic futures. It is a tool for exploring how sensitive a plan is, not an oracle.