The point of a forecast is not the number. It is the conversation about which assumptions move the number—and what would have to be true.
Claude can build the driver table, run the scenarios, and calculate the runway. You decide which assumptions deserve a range, which deserve a single number, and when the forecast should be reviewed again.
01
Hidden drivers
A single growth rate hides the assumptions that actually move the business.
02
Single-point certainty
One line implies a precision the model does not have.
03
Cash left out
Revenue scenarios that ignore burn and working capital can miss the real constraint.
02 THE PROMPT DECK
Forecasting prompts for Claude in Excel
Start with the driver table, then use the prompts to build the range, test the sensitivity, and translate the plan into cash.
01BUILD
Build a 12-Month Forecast
Build a 12-month revenue forecast from the historical data in this workbook. Explain the growth assumptions, seasonality, and any manual overrides. Keep formulas auditable and show the drivers on a separate tab.
Turn history into a driver-based forecast.
02SCENARIOS
Run Base, Upside, and Downside Scenarios
Create base, upside, and downside scenarios for this model. Change only the key drivers, keep formulas intact, and summarize the impact on revenue, EBITDA, and cash. Use a scenario switch instead of duplicating the workbook.
Compare three futures without copying the model.
03SENSITIVITY
Build a Sensitivity Table
Create a two-way sensitivity table for the selected model. Use the two most important assumptions and show the output range. Explain how to update it and add a note about the assumptions held constant.
Show how the answer moves across two assumptions.
04CASH
Forecast Cash Runway
Calculate cash runway under the current plan and two downside scenarios. Show the assumptions, monthly burn, and the month cash reaches zero. Explain which assumptions move the runway the most.
Translate the plan into months of cash.
05COHORTS
Forecast Cohort Retention
Project retention and revenue by cohort for the next 12 months. Use the historical cohort data and state the retention assumptions. Show the retention curve and the resulting revenue contribution.
Project revenue by cohort, not just totals.
06ASSUMPTIONS
Stress-Test Key Assumptions
List the five most optimistic assumptions in this model. For each one, show the cell reference, the current value, a conservative alternative, and the impact on revenue, EBITDA, and cash runway. Use a table.
It can help build or populate a model and explain the formulas. It is not a substitute for human review, especially for audit-critical or final deliverables.
How many scenarios should I run?+
Start with base, upside, and downside. More cases add complexity; the goal is a conversation about which assumptions matter.
Can Claude forecast cash runway?+
Yes. Give it cash, monthly burn, hiring plan, and any financing assumptions. Review the month cash reaches zero and the assumptions behind it.
Does Claude support data tables?+
No. Claude for Excel does not support data tables. Build sensitivity grids with formulas and review them manually.
How do I avoid false precision?+
Show a range, label every assumption, include a sensitivity table, and set a review date. A forecast is a decision tool, not a promise.
Can I use this for a startup board deck?+
Yes, as a drafting and analysis aid. Review the model, the assumptions, and the narrative before sharing.
Start with the six prompts above, then explore the full library for formulas, dashboards, audits, and forecasts. Download the PDF for a printable field guide.