AI Value
Realization Lab.
Separate productivity improvements from workflow economics, capacity capture, and financial returns. Change the assumptions. See which conclusions hold.
Build the business case.
Start with a sample workflow, then replace the assumptions with your own.
01Baseline & productivityHow much work does the process require today?⌄
02AI workflow & qualityWhat human work remains after automation?⌄
03Capacity & financial realizationWhat will the organization actually capture?⌄
04Fully burdened AI program costBuild a defensible estimate, not just a token bill.⌄
Read the cost estimation guide ↗ for a worked ledger, allocation rules, and an explanation of which costs belong in these fields.
Calculating your scenario…
Inspect the mechanics.
Open the methodology to see what each number means and where the model can mislead.
View formulas & limitations ＋
Required successful cases: Annual target volume. Expected AI attempts = target volume ÷ successful-attempt rate. The model assumes independent attempts and a stable success rate; it does not simulate individual retries or changing task complexity.
Baseline human cost: Target volume × baseline minutes ÷ 60 × fully burdened hourly rate.
AI workflow human cost: (Expected attempts × review minutes + expected failed attempts × additional recovery minutes) ÷ 60 × hourly rate. This assumes review on every attempt and recovery work on unsuccessful attempts.
Capacity created: Baseline human hours minus AI workflow human hours. Negative values indicate additional human workload. Cash benefits are never inferred from negative capacity.
AI program cost: Expected attempts × (inference + data cost per attempt) + fixed annual operations + one-time implementation investment ÷ evaluation horizon.
Workflow unit cost: (Human workflow cost + AI program cost) ÷ successful annual cases. This is a resource-cost comparison, not proof of savings in the income statement.
Financial benefits: Positive released hours × fully burdened rate × capture percentage + separately substantiated incremental contribution. The capture percentage is an explicit assumption, not an estimate of automatically realized value.
Net annual financial contribution: Financial benefits minus annualized AI program cost. This compares incremental benefits against AI costs and avoids subtracting remaining labor a second time. The model assumes baseline labor spending otherwise continues.
Boundaries: No NPV, discounting, taxes, implementation ramp, demand elasticity, error-related liability, or changes in quality valuation. Do not enter contribution that already includes labor benefits captured elsewhere. The same activity must not be counted as both avoided labor cost and incremental contribution.