Credit Forecasting
and Budget Planning
Turn usage and rollout plans into a clearer credit purchase decision.
Use it before you have usage history
Size your initial purchase from the number of users in each persona and how each persona uses Cowork.
Use it once you have usage history, from pilot to expansion
- Start from real usage: your average credits per user per month
- Model growth: plan non-linear rollouts, such as a pilot of 200 users expanding to 1,000 in one jump, then add ongoing monthly growth
- Know your runway: how long your credits last and when to buy next
- Close the loop: PowerPoint for Finance and business partners, or an Excel model to customize
Enter your assumptions, get your forecast, then share it.
Scenario estimates only. Review all starting values and prices. Data stays in this browser session; exports can contain your inputs. Estimates are not invoices or spending limits.
Enter your assumptions
Four short sections: where you start, how you grow, which credits you hold, and what they cost. Results update as you type.
1. Starting point
The first forecast month defaults to next month; change it if your plan starts at a different time. Find credits per user at aka.ms/copilotdashboard → Consumption Dashboard → Explore service insights.
2. Credits and prices
Your starting prepaid credits are used first and carry forward; only demand beyond them is billed at the pay-as-you-go price.
3. Growth and planned schedule
Start with standard monthly rates. Then, optionally, use the planned schedule to override them for specific months.
Standard assumptions
Users grow by the monthly rate (rounded to whole people) until they reach the maximum allowed users. Usage per user compounds monthly starting in month 2.
Planned schedule (optional)
Use this to plan future expansion and purchases. Enter the total users you expect after a step-up, for example when a new department or region is enabled at once, and any credits you plan to purchase in a given month. Purchased credits are added on top of your starting credits and are available at the start of that month. Months you leave blank follow the standard assumptions above; months you fill in override them.
| Month | Date | Total users after step-up | Credits purchased |
|---|
Your forecast
See the credits you may need and their estimated cost. This is based on credits used, not necessarily what you pay during the period.
Explore your forecast
Saved plan omitted from the chart and monthly details because the periods or data sources differ, or the source is unknown. See Compare two plans.
Chart and table values are rounded; the monthly forecast CSV carries more precision. Imported actuals are separate and only appear in the demand view.
Monthly forecast details
Department forecast details
Forecast-only group estimates share the single plan's credits and budget. Costs are estimates, not allocated invoices.
Prepaid and pay-as-you-go details
Scheduled purchases are available at the start of the month. Depending on your selection, unused prepaid credits carry forward or reset each month.
How costs and savings are estimated
Estimated cost = prepaid credits used × your prepaid price + pay-as-you-go credits used × your pay-as-you-go price. To estimate the difference, we compare that with paying the pay-as-you-go price for every credit used. The result can be negative. In the comparison, starting credits and scheduled purchases are valued at your prepaid price, but starting credits may already be paid for. Neither amount is an invoice or a forecast of when you will pay. Taxes, fees, expiration and other contract terms are not included. The budget comparison uses the estimated cost of credits used, not the amount paid to purchase them.
Test a what-if
Optional.
Additional budget
Money you could still spend, on top of the credits you already hold (starting balance and planned purchases, which are used first). See how many users it can support, how long it lasts and what usage it allows.
Plan by department
Enter explicit, non-overlapping groups. All groups share the one prepaid pool, purchases and budget; no allocation is repeated.
Compare two plans
Save a plan as your starting point, then change the inputs to explore another plan. Compare the credits, costs and timing to see what your changes could mean. Your saved plan stays unchanged until you save it again.
Take this plan to your team
Download a one-page executive overview or an Excel workbook with live formulas. Amounts are estimates, not invoices.
Exports are created in your browser and can contain sensitive group totals or notes. Review them before sharing.
Import actuals and save or reopen a scenario
For actuals, use a CSV with the header month,credits,cost,active. These figures are shown on the chart and monthly table only; they are not included in forecast totals or exports.
Scenario JSON saves assumptions, explicit departments, checkpoints, chart and PowerPoint settings, comparison, date-attribution caveats and your decision note, not imported usage history or source identity. After reopening, comparison differences are unavailable until you save the current plan again. Reapply your data to include history in a presentation; remove explicit department assumptions first, since imported aggregate data cannot provide a split.
Open an earlier version 1 scenario
Version 1 files use a separate, earlier model and cannot be converted automatically. This opens that model without changing your current forecast.
From rollout to runway.
Explore how adoption, usage and a shared prepaid allocation shape a phased credit budget. Compare a saved baseline with an editable alternative.
Synthetic illustrative data only. No live Power BI connection. User-entered and imported data stays in this browser session; nothing is stored or sent by this page.
Estimated costs depend on assumptions and are not commitments. Do not enter sensitive information in a public or shared environment.
Two scenarios, one decision
Forecast-only totals across each scenario’s full horizon. Actuals are never blended into these totals.
Credit demand over time
Independent forecast lines and imported actuals; no actuals replace forecast values.
Credits per month. Vertical lines mark review points; the dotted line shows monthly prepaid credits when applicable.
Decision brief
Accessible monthly forecast and actual table
Explore a scenario range
Not a prediction or confidence interval. Low and high change every alternative phase’s monthly usage growth by the selected percentage points; base uses the alternative unchanged.
Shape the alternative
Edits recalculate the alternative; the saved baseline stays unchanged until you explicitly replace it.
Costs and timing
Enter prices and budget in the same currency. The budget covers the whole forecast. Current month 0 is before rollout. Separate review months with commas.
Phases
Starts must be strictly increasing, begin at month 1, and fall within the horizon. Growth is monthly, including month 1. A blank name becomes “Phase N”.
| Optional name | Start month | Active user growth / month % | Usage / active user growth / month % | Newly enabled at start | New cohort activation % |
|---|
The prior phase’s ending active users and usage carry into the next phase. Growth is applied first; new enabled users are then allocated by remaining eligibility and activated once.
Departments
Counts may be fractional in forecasts. The prepaid allocation is shared, not repeated for each department.
| Name | Eligible | Enabled | Active | Credits / active user / month |
|---|
Forecast department totals
Alternative, summed across the forecast horizon. Totals reconcile before display rounding.
Phase contribution waterfall
Final-month credits: initial monthly demand plus sequential phase contributions. Each contribution is the difference when that phase is added to the earlier phases; later phases are held at the preceding phase’s rates with no new cohort.
Actuals & exports
Import a local CSV for visual comparison, or export the independent monthly forecasts.
CSV format: first row exactly month,credits,cost,active; one numeric row per month, with no extra columns. Month must be a unique integer from 1 to the alternative horizon; other values must be finite and nonnegative. Blank lines are ignored. No quoted values or formulas.
No actuals imported.