Free · ~5 minutes
Estimate the labour-cost side of an AI business case with honest assumptions. Build cost, run cost, realised savings (with the honesty discount most calculators skip), payback period, 3-year cumulative. Calculated in real time as you change the inputs.
Inputs
Loaded hourly cost: $90.71/hour
Be honest. The 80/20 reality of AI automation: most workflows have a realistic 40–70% automation ceiling once you account for edge cases the operator still handles.
Foundation model API + vector DB + infrastructure + observability.
Year-1 net
$-106K
Realised annual saving $64K minus annual opex $110K minus year-1 build cost $60K.
Payback period
—
Annual opex exceeds saving — no payback at current assumptions
3-year cumulative
$-198K
Year 1 + Year 2 + Year 3 net (build cost paid year 1 only)
How the math breaks down
What this isn’t
This calculator estimates the labour-cost side of the business case only. Revenue impact, risk reduction, and strategic capability gains aren’t modelled here — for a complete business case, see Measuring AI ROI.
Want a senior partner to walk through these numbers with you and put them inside a full four-category business case?
Most AI ROI calculators show gross savings: "this system will save 4 FTE worth of work per year — that’s $400K." A CFO sees through it in 10 seconds. The right question is: will those 4 FTE actually be eliminated, redeployed to higher-value work, or absorbed by previously-deferred work?
The "honesty discount" dropdown above is the field most other calculators don’t have. It forces the question. Picking 100% means a real CFO-visible saving — a role eliminated, a hire avoided, a contract cancelled. Picking 50% means the capacity is partially redeployed. Picking 0% means the time savings are real but won’t show up on the P&L.
Honest answer is usually somewhere between 25% and 75% for most AU mid-market AI investments. Almost never 100% — that’s only true for hire-avoidance scenarios where the headcount line item gets smaller.
This is the labour-cost side only. Revenue impact (new conversion, expanded sales capacity), risk reduction (fraud caught, compliance exposure avoided), and strategic capability (work that wasn’t possible at any cost before) all sit outside this calculator’s scope. For a complete business case, see the CFO-grade framework cornerstone.
The default annual ops-people cost of $50K is a realistic AU mid-tier estimate — typically 1–4 days/month of senior engineering maintenance, plus 4–8 hours/week of internal AI-champion time. For lighter or heavier ops models the number shifts; treat it as an order-of-magnitude figure rather than a precise forecast.
The output is a starting point for a CFO conversation, not a final business case. The reliable next step is to take the calculator’s structured output, plug your specific workflow into it, sense-check the assumptions with your operations team, and produce a one-page business case for your finance partner. The four-category model in Measuring AI ROI is the template most CFOs respond to.
For engagements where the math suggests pursuit, the productised AI Readiness Sprint is the 2-week engagement that takes the calculator output, validates it against your specific data and operations reality, and produces a written go/no-go recommendation with the criteria authored upfront.