Workshop / Legal AI

Does the pilot deserve to scale?

See how review effort changes the business case and whether the pilot meets its quality requirements.

← Investment portfolio
Your assignment: challenge a promising pilot

Start with the base case. Increase review time, test hourly billing, then inspect what would change the decision.

$50,870 / 12 moReview candidate

Assumptions

Volume and eligibility

Every attempted item incurs preparation and review. Rejected outputs then require a full manual redo. A unit is one bounded item, such as an agreement comparison.

Ready for an expansion review. Approval still requires supporting evidence.
149.5hnet capacity / month780 attempted items
$0realizable cash / monthNot the value of salaried time
$50,870net over 12 monthsIncludes contribution and setup

Where the time goes each month

Existing workflow390.0 hPreparation-104.0 hHuman review-78.0 hManual redo-58.5 hCapacity released149.5 h
Existing workflow390.0 h
Preparation-104.0 h
Human review-78.0 h
Manual redo-58.5 h
Capacity released149.5 h
TotalBenefitWork or cost deducted

663 accepted + 117 rejected = 780 attempts. Verification is charged on every attempt.

Return to portfolio →

Assumptions, persistence and boundaries

All figures are fictional, not vendor prices or evidence about a real firm. Inputs persist in this browser tab when session storage is available. Draft inputs do not affect the portfolio until explicitly saved as an assessment. A stage or delivery change makes a saved assessment stale. Failed saved assessments hold expansion; discovery and bounded pilots remain separate decisions. Existing fixed fees stay unchanged. The model excludes ramp-up, taxes, discounting and additional staffing cost when capacity is negative.