Automation calculator
Let N be people, T tasks per person per month, M minutes per task, H hourly cost, R time reduction as a fraction, C cash conversion as a fraction, S monthly software and maintenance cost, and I initial setup cost.
- Baseline hours = N × T × M ÷ 60.
- Released hours = baseline hours × R.
- Capacity value = released hours × H.
- Realized cash savings = capacity value × C.
- Net monthly cash = realized cash savings − S.
- First-year net cash = 12 × net monthly cash − I.
- Simple payback = I ÷ positive net monthly cash.
If monthly net cash is not positive, the tool reports payback as not reached. Operating break-even excludes initial setup costs. If conversion or baseline value is zero while recurring cost is positive, operating break-even cannot be achieved within this model.
Workflow simulator
Manual minutes = tasks × manual minutes per task. Assisted minutes = tasks × (generation + review + error fraction × additional rework). The percentage reduction uses manual minutes as the denominator. A negative result indicates more effort than the manual baseline.
Example: 100 tasks, manual time 20, generation 2, review 5, error rate 10%, rework 10 gives 2,000 manual minutes and 800 assisted minutes: 1,200 minutes released, or 60%. The stage animation shows sequence, not elapsed processing time.
Prompt Lab
The builder inserts your entered task, audience, context and constraints into a fixed template, then adds verification instructions. It performs no external research, does not check supplied claims and does not call an AI model. Generated text is displayed as text rather than executed as HTML.
Validation and limits
Development checks covered the calculator worked example, zero savings and workload, simulator arithmetic including a negative-benefit case, and browser operation of the simulator and prompt builder. This is software verification, not independent validation of an economic forecast. Your inputs determine whether the result is useful.
The tools exclude taxes, discounting, ramp-up, parallel scheduling and consequential effects of mistakes. Use them to explore assumptions, then validate those assumptions through a measured pilot. Rounded display values can differ slightly from exported unrounded values.