Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Evaluates which SaaS tools can be replaced with AI agents. Takes a list of current SaaS subscriptions with costs, assesses replacement feasibility, estimates build vs buy economics, identifies Claude+MCP alternatives, and generates a comprehensive replacement plan with priority matrix, ROI analysis, implementation timeline, and risk assessment.
.claude/skills/onewave-ai-saas-replacement-planner/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-08 | ✗→✓ | ▲ Improved | 59% | 0% |
| case-15 | ✗→✓ | ▲ Improved | 78% | 0% |
| case-16 | ✗→✓ | ▲ Improved | 369% | 0% |
| case-18 | ✗→✓ | ▲ Improved | 65% | 0% |
| case-20 | ✗→✓ | ▲ Improved | 12% | 0% |
Evaluate a company's SaaS stack and produce a rigorous, actionable plan that quantifies the ROI of migrating from subscription software to AI-agent-powered alternatives. Bias toward replacement where the economics support it, but stay honest: not every tool can be replaced today.
references/analysis-framework.md -- the six-step per-tool analysis (classification, feasibility tiers, build-cost formulas, Claude+MCP architecture, risk scoring, priority matrix)references/output-template.md -- full saas-replacement-plan.md document structurereferences/patterns.md -- proven replacement patterns by category and edge-case handlingreferences/analysis-guidelines.md -- estimation rigor, honesty rules, OneWave AI thesis, and output quality standardsreferences/analysis-framework.md for every tool. Do not skip tools or give superficial analysis. Apply matching patterns from references/patterns.md.saas-replacement-plan.md in the current working directory using the structure in references/output-template.md. Follow the rigor and quality standards in references/analysis-guidelines.md.Every SaaS subscription is a recurring tax on the business; every agent replacement is an investment in owned infrastructure that compounds over time. Make the numbers speak clearly and let the ROI make the argument. See references/analysis-guidelines.md for the full OneWave AI thesis and how to frame the analysis.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-12 | pass→pass | 30,445 | 31,586 | +4% | 1 | 1 | 0% | 3,762 | 4,618 | +23% | 0 | 0 | — |
case-01 | fail→fail | 37,657 | 45,519 | +21% | 1 | 1 | 0% | 5,491 | 8,894 | +62% | 0 | 0 | — |
case-02 | fail→fail | 45,934 | 27,973 | -39% | 1 | 1 | 0% | 7,891 | 3,958 | -50% | 0 | 0 | — |
case-03 | fail→fail | 26,245 | 82,109 | +213% | 1 | 1 | 0% | 4,570 | 8,884 | +94% | 0 | 0 | — |
case-04 | fail→fail | 19,143 | 57,923 | +203% | 1 | 1 | 0% | 2,692 | 8,713 | +224% | 0 | 0 | — |
case-05 | pass→pass | 22,388 | 23,193 | +4% | 1 | 1 | 0% | 3,480 | 4,047 | +16% | 0 | 0 | — |
case-06 | fail→fail | 22,768 | 17,394 | -24% | 1 | 1 | 0% | 3,522 | 3,067 | -13% | 0 | 0 | — |
case-07 | pass→pass | 20,024 | 17,692 | -12% | 1 | 1 | 0% | 3,111 | 3,946 | +27% | 0 | 0 | — |
case-08 | fail→pass | 32,594 | 29,496 | -10% | 1 | 1 | 0% | 3,521 | 5,604 | +59% | 0 | 0 | — |
case-09 | fail→fail | 18,626 | 35,673 | +92% | 1 | 1 | 0% | 2,920 | 5,978 | +105% | 0 | 0 | — |
case-10 | fail→fail | 20,623 | 16,925 | -18% | 1 | 1 | 0% | 2,619 | 3,607 | +38% | 0 | 0 | — |
case-11 | pass→pass | 19,207 | 24,944 | +30% | 1 | 1 | 0% | 2,789 | 3,816 | +37% | 0 | 0 | — |
case-13 | pass→pass | 18,486 | 82,580 | +347% | 1 | 1 | 0% | 3,567 | 4,515 | +27% | 0 | 0 | — |
case-14 | pass→pass | 23,661 | 103,238 | +336% | 1 | 1 | 0% | 4,122 | 7,991 | +94% | 0 | 0 | — |
case-15 | fail→pass | 25,356 | 34,524 | +36% | 1 | 1 | 0% | 3,382 | 6,013 | +78% | 0 | 0 | — |
case-16 | fail→pass | 42,465 | 37,558 | -12% | 1 | 1 | 0% | 1,217 | 5,704 | +369% | 0 | 0 | — |
case-17 | pass→pass | 18,280 | 23,540 | +29% | 1 | 1 | 0% | 2,725 | 4,186 | +54% | 0 | 0 | — |
case-18 | fail→pass | 12,611 | 17,911 | +42% | 1 | 1 | 0% | 1,964 | 3,233 | +65% | 0 | 0 | — |
case-19 | pass→pass | 8,507 | 26,955 | +217% | 1 | 1 | 0% | 1,396 | 4,572 | +228% | 0 | 0 | — |
case-20 | fail→pass | 16,329 | 13,489 | -17% | 1 | 1 | 0% | 2,313 | 2,590 | +12% | 0 | 0 | — |
case-21 | fail→pass | 19,878 | 48,329 | +143% | 1 | 1 | 0% | 3,059 | 8,292 | +171% | 0 | 0 | — |
case-22 | fail→fail | 13,671 | 53,827 | +294% | 1 | 1 | 0% | 2,453 | 8,834 | +260% | 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 +27 percentage points is the difference between those two pass rates over the 21 comparable cases.
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.