A calculator that turns AI adoption inputs into hours saved, cost avoided and a payback period — using only the numbers you supply. It is the practical companion to Making the Financial Case for AI.

Its purpose is to make the arithmetic visible. A case built on a headline saving invites a search for the assumption that produced it; a case built on visible inputs survives the search.

You cannot fund a fix for a cost you cannot see. Start by measuring the hours.

What it calculates

You enter six inputs for one workflow, and the calculator returns five outputs.

Inputs

  • Cycles per year — how often the workflow runs.
  • Hours per cycle today — end to end, including review.
  • Hours per cycle after — including the verification step you are adding.
  • Loaded hourly cost — salary plus employer costs and overhead, as finance accounts for it.
  • Adoption factor — the share of cycles that will actually run through the new workflow in year one.
  • Setup and training hours — the one-off effort, valued at the same loaded rate.

Outputs

  • Hours returned per cycle — the difference between before and after, times the adoption factor.
  • Annual hours returned — hours per cycle times cycles per year.
  • Annual cost avoided — annual hours at the loaded rate.
  • Year-one net benefit — annual cost avoided, less the setup and training cost and any tool subscription.
  • Payback period — setup and tool cost divided by the annual benefit.

A worked example (illustrative)

The figures below are illustrative, to show how the calculator works. Substitute your own; every input is yours, and no benchmark is used.

  • Cycles per year: 12 (monthly client reports).
  • Hours per cycle today: 14.
  • Hours per cycle after: 10.5 (including verification).
  • Loaded hourly cost: $65.
  • Adoption factor: 80%.
  • Setup and training hours: 20, plus $600 in tool seats.

Calculation:

  • Hours returned per cycle: (14 − 10.5) × 0.8 = 2.8 hours.
  • Annual hours returned: 2.8 × 12 = 33.6 hours.
  • Annual cost avoided: 33.6 × $65 = $2,184.
  • Setup and tool cost: (20 × $65) + $600 = $1,900.
  • Year-one net benefit: $2,184 − $1,900 = $284.
  • Payback: roughly 11 months.

The number is modest, and that is the honest output for a single report type. The case becomes compelling when the same workflow is applied to three or four recurring documents, because the setup is largely shared — which is why the portfolio view matters more than the single-use-case view.

How to use it

  • Count the whole cycle, including the verification you are adding. A calculator that omits the check will overstate the benefit.
  • Use a loaded hourly cost, so the figure matches how finance accounts for time.
  • Be conservative on the adoption factor. Real adoption ramps, and a workflow nobody uses returns nothing.
  • Include the shared setup once. The second workflow reuses most of the first one’s design.
  • Re-run after two cycles, with the actual hours, and compare against this estimate.

What the number is for

The figure does three jobs.

  • It makes the cost visible, which is the precondition for funding a change.
  • It prioritizes the workflow list, because the most expensive recurring work is the best candidate.
  • It sets a baseline, so the improvement can be measured rather than asserted.

Where the number is large, the case is usually self-evident. Where it is small, the honest answer may be that this workflow is not worth changing — which is also a useful result.

Frequently asked questions

How do you calculate the ROI of AI adoption?

Multiply the hours returned per cycle by the cycles per year and the loaded hourly rate, subtract the setup, training and tool costs, and divide the setup investment by the annual benefit to get the payback period.

What inputs do I need?

Cycles per year, hours per cycle before and after, loaded hourly cost, an adoption factor, and the one-off setup and training hours.

Should I include the verification time?

Yes. Verification is real time every cycle, and a case that omits it will overstate the benefit — sometimes to the point of making an unviable workflow look viable.

What adoption factor should I use?

A conservative one. Eight in ten cycles is a reasonable first-year assumption for a well-designed workflow; higher figures should be justified rather than assumed.

Does the calculator use industry benchmarks?

No. It uses only the inputs you supply, so the result is your cost rather than an average.

How do I know whether the case was right?

Re-run the calculation after two cycles with the actual hours, and compare. A case that cannot be tested afterwards was an estimate, not a case.

Next step

Run your own numbers for one workflow, then re-run them after two cycles. See Making the Financial Case for AI and Measuring AI Adoption and Return, or book an AI adoption call and we will baseline your cycle with you.

Sources

  • The calculator performs arithmetic on user-supplied inputs only. It uses no benchmark data and makes no claim about typical savings.

The worked example is explicitly illustrative. Where third-party figures appear in the linked guides, they are cited there with their source and date.