Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Calculate comprehensive ROI for AI implementation projects. Takes current costs, manual process time, team size, and hourly rates. Generates detailed roi-analysis.md with executive summary, cost-benefit tables, sensitivity analysis, break-even timeline, and comparison scenarios. Use when evaluating AI investments, building business cases, or justifying automation spend.
.claude/skills/onewave-ai-roi-calculator/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-09 | ✗→✓ | ▲ Improved | 103% | 0% |
| case-12 | ✗→✓ | ▲ Improved | 204% | 0% |
| case-17 | ✗→✓ | ▲ Improved | 148% | 0% |
| case-06 | ✓→✗ | ▼ Worse | 167% | 0% |
| case-20 | ✓→✓ | ＝ Same ✓ | 183% | 0% |
Gather inputs about current operations and produce a comprehensive roi-analysis.md with data-backed financial insights and a clear recommendation.
references/inputs.md — required and optional inputs, with defaultsreferences/methodology.md — all calculation formulas and sensitivity/comparison logicreferences/output-template.md — full roi-analysis.md structure to fill inreferences/rules-and-protocol.md — calculation rules, interaction protocol, quality checklistreferences/inputs.md for the full input list and defaults.references/inputs.md and note that defaults were applied.references/methodology.md.references/methodology.md.roi-analysis.md to the current working directory following references/output-template.md. Apply the formatting and calculation rules in references/rules-and-protocol.md.references/rules-and-protocol.md, then report the top 3 findings and the saved file path.| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 38,175 | 27,206 | -29% | 1 | 1 | 0% | 8,303 | 729 | -91% | 0 | 0 | — |
case-02 | fail→fail | 19,331 | 43,221 | +124% | 1 | 1 | 0% | 3,638 | 8,656 | +138% | 0 | 0 | — |
case-03 | fail→fail | 30,753 | 41,713 | +36% | 1 | 1 | 0% | 6,602 | 8,682 | +32% | 0 | 0 | — |
case-04 | fail→fail | 14,056 | 40,774 | +190% | 1 | 1 | 0% | 2,901 | 8,624 | +197% | 0 | 0 | — |
case-05 | fail→fail | 18,417 | 28,816 | +56% | 1 | 1 | 0% | 3,703 | 6,821 | +84% | 0 | 0 | — |
case-06 | pass→fail | 16,928 | 38,658 | +128% | 1 | 1 | 0% | 3,234 | 8,629 | +167% | 0 | 0 | — |
case-07 | fail→fail | 16,916 | 36,956 | +118% | 1 | 1 | 0% | 3,364 | 8,620 | +156% | 0 | 0 | — |
case-08 | fail→fail | 20,523 | 41,669 | +103% | 1 | 1 | 0% | 3,908 | 8,620 | +121% | 0 | 0 | — |
case-09 | fail→pass | 21,544 | 38,975 | +81% | 1 | 1 | 0% | 4,241 | 8,614 | +103% | 0 | 0 | — |
case-10 | fail→fail | 14,459 | 99,214 | +586% | 1 | 1 | 0% | 2,924 | 8,620 | +195% | 0 | 0 | — |
case-11 | fail→fail | 18,180 | 39,958 | +120% | 1 | 1 | 0% | 3,627 | 8,621 | +138% | 0 | 0 | — |
case-12 | fail→pass | 14,345 | 38,076 | +165% | 1 | 1 | 0% | 2,834 | 8,607 | +204% | 0 | 0 | — |
case-13 | fail→fail | 14,892 | 49,276 | +231% | 1 | 1 | 0% | 3,091 | 8,619 | +179% | 0 | 0 | — |
case-14 | fail→fail | 15,715 | 40,254 | +156% | 1 | 1 | 0% | 3,409 | 8,620 | +153% | 0 | 0 | — |
case-15 | fail→fail | 13,839 | 38,624 | +179% | 1 | 1 | 0% | 2,687 | 8,606 | +220% | 0 | 0 | — |
case-16 | fail→fail | 16,975 | 39,903 | +135% | 1 | 1 | 0% | 3,329 | 8,622 | +159% | 0 | 0 | — |
case-17 | fail→pass | 15,423 | 94,091 | +510% | 1 | 1 | 0% | 3,473 | 8,619 | +148% | 0 | 0 | — |
case-18 | fail→fail | 15,144 | 38,524 | +154% | 1 | 1 | 0% | 2,821 | 8,619 | +206% | 0 | 0 | — |
case-19 | fail→fail | 27,489 | 38,908 | +42% | 1 | 1 | 0% | 5,850 | 8,619 | +47% | 0 | 0 | — |
case-20 | pass→pass | 13,676 | 38,939 | +185% | 1 | 1 | 0% | 3,050 | 8,620 | +183% | 0 | 0 | — |
case-21 | pass→pass | 10,176 | 28,669 | +182% | 1 | 1 | 0% | 2,001 | 5,691 | +184% | 0 | 0 | — |
case-22 | pass→pass | 11,474 | 23,672 | +106% | 1 | 1 | 0% | 2,656 | 5,874 | +121% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted, and 21 counted toward the lift figure. The other 1 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of +9 percentage points is the difference between those two pass rates over the 21 comparable cases. 2 cases got worse with the skill loaded, and they are included in that figure.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.