A decision model designed to resist optimistic arithmetic.
The audit separates cash savings from capacity, prevents common double-counting and leads with a downside case.
What the instrument reports
Two headline scores remain separate: Value estimates economic strength; Coordination estimates whether the organization can reliably realize it. Risk flags can veto an expansion recommendation, but they never average into a reassuring score.
Core formulas
Hard value = monthly costs actually eliminated × 12.
Capacity value = weekly hours saved × 52 × hour value × usable-capacity rate × adoption.
Attributed margin value = influenced revenue × gross margin × defensible attribution × adoption.
Expected ROI = (annual defensible value − annual modeled cost) ÷ annual modeled cost.
When saved capacity also produces the reported revenue, the model uses the larger of those two values instead of adding both. This blocks the same hour from being valued twice.
Scenario design
Known costs remain fixed. The conservative floor reduces adoption, discounts capacity, halves revenue attribution and excludes avoided-risk estimates. The expected case uses entered assumptions. The optimistic case is explicitly labeled as a ceiling and increases only uncertain benefit inputs.
Worked example
A $500 monthly tool with $2,500 of setup cost replaces $300 in monthly software. It saves eight hours weekly at $55 per hour, but only 60% of that capacity is usable and adoption is 55%. Revenue influence is adjusted by gross margin and attribution. If those hours also created that revenue, only the larger value is counted. The recommendation then considers break-even adoption, coordination readiness, tenure and unresolved veto risks.
Limitations
This is decision support, not a financial audit, valuation, legal review or security assessment. Capacity, attribution and avoided-risk estimates are subjective. Validate the result against invoices, payroll decisions, workflow evidence and realized gross margin before committing capital.
Privacy design
The public audit is deliberately ephemeral. Inputs are calculated in the page and are not transmitted, saved to browser storage or placed in a database.
Changelog
9Q-AIA 1.0 · July 26, 2026 — Introduced hard/soft separation, double-count protection, tenure gates, floor-first scenarios, break-even adoption and risk vetoes.
Run the AI Investment Audit