Executive summary. Across 20,000 seeded simulations, this investment returns a positive ROI in 71.4% of outcomes, with a median NPV of over 3 years (poor case , strong case ). The verdict is CONDITIONAL-GO: the economics clear the hurdle in the base case, but the margin of safety is thin. Two variables carry most of the downside — adoption rate and whether the workflow is genuinely redesigned rather than merely tooled.
The team spends roughly 3,220 hours a year on this workflow. At the loaded rate provided, the status quo consumes $•••,••• annually, or $•••,••• across the 3 year horizon. That figure is the true baseline the investment is measured against — not zero.
| Hours spent on this task per year | 3,220 | Team size × hours per week × working weeks |
| Annual labor cost of the task today | $•••,••• | The savings pool the investment draws from |
| Upfront after expected overrun | $•••,••• | Implementation costs typically land above the quote |
| Total cost in present value | $•••,••• | Discounted at your hurdle rate |
| Expected time reduction on the task | 24.6% | Before adoption and quality effects |
| Expected adoption | 55.9% | Share of the team actually using it day to day |
| Effective capture rate | 13.7% | Time saved × adoption × quality — what reaches the P&L |
| Hours freed per year (expected) | 441 | Success-weighted, at full ramp |
| Net annual savings (expected) | $••,••• | After recurring cost |
| Expected time to first value | 10 months | Ramp before savings appear |
| IRR (expected case) | 38.2% | Return rate implied by expected cash flows |
| Benefit-cost ratio | 1.84× | Value returned per dollar committed |
What it must clear to be worth doing
| Annualized cost to beat | $••,••• | Upfront spread over the horizon, plus recurring |
| Hours that must be freed per year | 612 | At the loaded hourly cost entered |
| Hours per person per week | 1.9 | The break-even per individual — sanity-check this with the team |
| Cost per person per year | $•,••• | Annualized total cost ÷ team size |
| Bottom 5% (P5) | Negative | 1 in 20 outcomes are worse than this |
| Lower quartile (P25) | $••,••• | 1 in 4 outcomes are worse |
| Upper quartile (P75) | $•••,••• | 1 in 4 outcomes are better |
| Top 5% (P95) | $•••,••• | Realistic ceiling, not a promise |
| Average of the worst 10% | -$••,••• | The loss you should be able to absorb |
| Chance of losing more than you invest | 4.1% | Downside worse than total committed cost |
Cumulative net cash turns positive in month 14 in the median case, month 9 in the strong case, and never within the horizon in the poor 10% of runs. Payback later than month 18 is the single clearest early warning that adoption is lagging the plan.
The frontier crosses break-even at roughly 46% adoption paired with a 21% time reduction. Anything below that pairing turns the verdict negative regardless of how well the vendor performs. Above it, returns compound quickly — the upper right quadrant of the frontier produces the strong case outcomes.
| Scenario | Adoption | Time saved | NPV (3yr) | Payback |
|---|---|---|---|---|
| Poor case (P10) | 22% | 12% | Negative | Never |
| Conservative (P25) | 35% | 18% | ••••• | Month 22 |
| Base case (P50) | 48% | 25% | ••••• | Month 14 |
| Upside (P75) | 62% | 33% | ••••• | Month 11 |
| Strong case (P90) | 74% | 41% | ••••• | Month 9 |
| Lever | Change | NPV impact | How you measure it |
|---|---|---|---|
| Adoption rate | +10 pts | High | Weekly active users vs. licensed seats |
| Time saved per task | +5 pts | High | Measured handle time, before vs. after |
| Recurring annual cost | −10% | Medium | All-in invoice at projected volume |
| Ramp length | −1 month | Medium | Weeks to first measured saving |
| Rework rate | −5 pts | Low | Share of output needing human correction |
Recommendation: CONDITIONAL-GO. The proposal clears its hurdle rate in 71.4% of modeled outcomes with a median 3 year NPV of $•••,•••, against a status quo cost of $•••,•••. The margin of safety is thin rather than comfortable, and the downside is concentrated in two controllable variables: adoption and workflow redesign. We recommend approving a capped 90 day pilot with an explicit exit, rather than the full rollout, and revisiting the full commitment against measured adoption and time saved at day 90.