Built by the platform from a two-paragraph brief, start to finish. A single-site, 3-month pilot, scoped to answer one question before the larger seven-site program gets funded: does it actually work.
One case, eight linked stages, from intake to the ask. Nothing here is re-keyed between steps, and sized right for a fast go/no-go, not the full program's case.
No new architecture bets. The pilot reuses the platform and integration pattern already designed for the full program, so the only open question is the one the pilot exists to answer.
The ask, the return, and the recommendation, with the confidence level stated, not implied. A 12-month payback, sized to get a real answer without betting the full program.
Move the sliders and ROI, NPV, IRR, and payback all re-price together, not just one headline number, within the engine's confidence band.
The platform scored the obvious shortcut, skipping the pilot and committing straight to the full program, and ranked it below the pilot on cost, time to value, and risk. You see the trade-off, not just the winner.
Even a clean, low-risk pilot gets this: the case flags that one site is not yet program-wide proof, and where an assumption is still unmeasured. A committee will ask these questions. This gets there first.
Bring a brief. In one session, we'll turn it into a case like this one.