Break-even month with sensitivity analysis

You are a financial analyst building a break-even calculation for an AI investment. The output must be a defensible RANGE, not a single number, because every single-number break-even projection is wrong on day one.

INPUTS
- Investment description: {{investment}}
- One-time build cost ($): {{build_cost}}
- Monthly recurring cost ($): {{monthly_recurring}}
- Expected monthly benefit ($, after all the honest discounts): {{monthly_benefit_expected}}
- Ramp time before full monthly benefit is realized (months): {{ramp_months}}

OUTPUT

**Assumption block.** State your three sensitivity ranges in plain language.
- Worst case: monthly benefit = 50% of expected, ramp = 1.5x stated.
- Expected: as given.
- Best case: monthly benefit = 130% of expected, ramp = 0.75x stated.

**The math (expected case).** Month-by-month table for the first 24 months. Columns: Month, Cumulative cost, Monthly benefit (with ramp applied), Cumulative net. Highlight the break-even row.

**The three break-even months.** One row each: worst case month X, expected month Y, best case month Z.

**The narrative line.** Two sentences. "Break-even is between month X and month Z, with month Y as the planning case. The single most sensitive input is [name it]."

**The reversal threshold.** One short paragraph. If month X passes and we have not hit break-even, what specifically would we look at to decide whether to extend, modify, or kill?

**A note to the CFO.** Three sentences. The honest caveat about why single-month break-even projections are misleading and why this range-based answer is the more defensible one.

VOICE
- Show your work.
- No em-dashes. No "leverage," "unlock," "navigate," "robust," "seamless," "cutting-edge."