AI VALUE AND COST · Calculator
AI Project ROI & Payback
Project the financial benefits and full cost of an AI workflow. Keep productivity value separate, then compare the plan with realized results in the toolkit workbook.
- Runs in your browser
- No sign-in
- Result before any contact step
- Capacity view
- +$58,000 net · ROI 152.63% · payback month 4
- Cash view
- −$14,000 net · ROI −36.84% · payback ≈ month 40 (estimate)
Capacity view pays back in month 4; the cash view has a monthly net of $500 and recovers setup around month 40.
This calculator produces planning estimates. It does not establish realized benefits. Compare measured results with the same period and scope using the AI Investment and Business Case Toolkit.
Enter the task, the people, the costs and the span. The calculator shows both views side by side, with payback, break-even adoption and a low, base and high range, so the case can be argued from assumptions that are written down.
Your inputs · worked out in your browser
0 of 8 required fields filled · empty
Waiting for inputs
Result
Enter the task, people, cost and horizon numbers, or load the example, to see capacity value and cost savings side by side.
What you get
- Hours freed, capacity value and cost savings, shown as separate figures
- Return and payback for the projected productivity value and for the projected financial benefits, in words
- Break-even adoption, flagged when it is above a full rollout
- Low, base and high cases with the printed illustrative factors, and an optional Scenario B
- A checklist of evidence to gather before the numbers are shared
If you accept analytics cookies, we record only which tool was opened, that a result was shown and which export or service link was used. Inputs, answers and results are never part of that record.
Method version 2026-10-14. Planning estimate from your own assumptions. It is not a forecast, a quote or financial advice, and it does not promise any saving or return.
How it works
Hours freed in a month equal tasks, times the minutes saved per task, divided by 60, times adoption. Count everything a person still does with the AI inside the minutes with the AI: writing the request, checking the output and fixing mistakes. If adoption ramps, it rises in equal steps from month 1 to the month of full adoption.
The projected productivity value prices the freed hours at the loaded hourly cost you enter and adds any other monthly profit contribution. The projected financial benefits counts only the share of those hours that becomes real spending cuts, plus the same other benefit. Cost savings are part of the capacity value, so the two figures are shown apart and are never added together.
In each view, net benefit is the benefit minus total cost, and total cost is the one-time setup plus the monthly running cost for every month of the span. Return on investment is net benefit divided by total cost. When the total cost is zero, the return is not defined and the page says so.
Payback is the first month in which the benefit so far, less the setup and the running costs so far, is zero or more. If that happens after your span, the page names the estimated month and says it is later. If the monthly net is zero or below, payback is never reached at full adoption. If it is more than 100 years away, the page says that instead of a month.
Break-even adoption is the share of tasks that must use the AI for the monthly financial benefit to cover the monthly running cost. It is flagged when it is above 100%, because no amount of rollout then closes the gap. The low and high cases apply the printed factors to your numbers. They are illustrative and are not forecasts.
Formula and rules
minutes saved = minutes before − minutes with the AI (checking included) hours freed in month m = tasks × minutes saved ÷ 60 × adoption × ramp(m) ramp(m) = the smaller of 1 and m ÷ months to full adoption (1 when the ramp is 0) capacity value = hours freed × loaded hourly cost cost savings = capacity value × cost-savings realization (part of the capacity value) capacity benefit = capacity value + other monthly profit contribution financial benefit = cost savings + other monthly profit contribution total cost = setup + monthly running cost × months in the span net benefit = benefit − total cost; ROI = net benefit ÷ total cost payback = first month where benefit so far − setup − running costs so far is zero or more break-even adoption = (running cost − other benefit) ÷ cost savings at full adoption illustrative low case: minutes saved × 0.75, adoption × 0.75, setup and running costs × 1.20 illustrative base case: exactly what you typed illustrative high case: minutes saved × 1.25, adoption × 1.25 (never above full adoption), costs × 1.00
Assumptions
- Every task takes the same time. The tool works with averages, not the spread between easy and hard tasks.
- Freed hours are priced at the loaded hourly cost you type. That measures capacity. It is not money in a bank account.
- The other monthly profit contribution counts in full from month 1. It does not ramp.
- The factors 0.75, 1.25 and 1.20 are illustrative. Replace your base numbers with ranges you can defend rather than relying on them.
- Costs you leave out, such as training, rework, extra review or human oversight, make the result look better than it is.
Worked example
Illustrative worked example · example valuesSample
A team handles 2,000 tasks a month. Each takes 12 minutes now and 4 minutes with the AI, checking included. The loaded cost is $50 an hour, 60% of tasks use the AI and use starts in month 1. That frees 160 hours a month, worth $8,000 of capacity in each month.
The example assumes that 25% of the labor valuation becomes cost savings, through less overtime and fewer contractor hours: $2,000 in each month. Setup costs $20,000 and running costs come to $1,500 in each month, so a 12-month span costs $38,000. Projected productivity value: $96,000 of benefit less $38,000 is $58,000 net, a valuation ratio of 152.63% (not financial ROI), paid back in month 4. Projected financial benefits: $24,000 of benefit less $38,000 is -$14,000 net, an ROI of -36.84%.
The projected financial case does not recover setup costs within 12 months. It nets $500 in each month, so the $20,000 setup is recovered around month 40, which is an estimate. Break-even adoption is 45%: below that share of tasks, the monthly financial benefit does not cover the running cost. The projected productivity value only holds if the freed time is worth its full hourly cost to you.
Evidence to gather
Records that make the result easier to check and act on:
- A time study of real tasks, before and with the AI, with checking and rework included.
- Monthly task counts from the system of record for the last 6 to 12 months, including seasonal peaks.
- Usage logs from a pilot or a comparable rollout that show real adoption and how quickly it grew.
- The loaded hourly cost from finance, and the budget line where any cost saving would appear.
- Quotes and contract terms for licences, usage pricing at your volume, integration, review, training, monitoring and oversight.
- A date to compare actual minutes, adoption, cost and expense reductions with the estimate.
What it does not do
- The result uses only the numbers you type. It does not know your team, your data or what the AI will really do.
- Time freed is not money saved. Only the share you enter as cost-savings realization counts as a spending cut.
- The low and high cases use fixed illustrative factors. They show sensitivity, not a forecast.
- It is a planning estimate, not a quote, financial advice or a statement about risk, legal or regulatory standing.
Questions
What is the difference between capacity value and cost savings?
Capacity value is the time freed, priced at your loaded hourly cost. Cost savings are the part of that time that becomes real spending cuts, such as less overtime, fewer contractor hours or a hire not made. Cost savings sit inside the capacity value, so the two are never added together.
Why are projected financial benefits often much lower than the projected productivity value?
Pay does not fall when a task gets faster, and freed time usually goes into other work. That can still be worth having if the time goes to work that earns money or clears a backlog. If you can support such a gain, enter it as other monthly profit contribution.
What should I count as minutes with the AI?
Everything a person still does for the task: writing the request, checking the output, fixing mistakes and handling the cases the AI cannot do. If you leave out checking time, the result will look better than it should.
What is a loaded hourly cost?
The full cost of one hour of the people doing the task: pay plus payroll taxes, benefits and the usual overhead. Ask finance for it. The plain wage understates the cost and a billing rate overstates it.
Are the low and high cases forecasts?
No. They apply the printed factors to your base numbers so you can see how far the answer moves. The factors are illustrative guesses. The stronger approach is to enter a range you can defend as your base and compare it with a Scenario B.
How should oversight and governance costs be handled?
Put them in the running cost: monitoring, human review, testing, vendor management and incident handling. A project that needs a person to approve every output has a different cost and a different minutes-with-the-AI figure than one that does not.