Library › Book 1 › Decision Framework -- When to Use What
Decision Framework -- When to Use What
23.1🔒Quick Decision TreeChapter 22 gave you the matrix; this chapter gives you the algorithm:
23.2🔒Scenario-Based RecommendationsContext: Small team, modern SPA, GitHub Actions CI, a coding agent (Claude Code, Cursor) already in the workflow.
23.3🔒Cost Reasoning at Scale: Ratios over Dollar TablesEarlier editions printed per-test dollar figures -- accurate when printed, misleading within months. The durable version:
23.4🔒Migration PathsAll three paths share a property: no phase requires a big-bang rewrite, and every phase produces value even if you stop there.
23.5🔒Risk Assessment: Plans Outlive ToolsOctomind -- credible and recommended six months ago -- is winding down. The defense is architectural:
23.6🔒The Recommended Stack (July 2026)- Primary: Playwright CLI + skill for agent-driven execution; Test Agents (planner/generator/healer) for the workflow - Secondary…
23.7🔒Key Takeaways- The first decision question is filesystem access: it determines whether the ~4x-cheaper CLI path is available at all. - Match tools to…
23.8🔒Q&ASelf-Assessment Quiz1. What is the first question once an AI agent drives the automation, and why? 2. Why does this book refuse to print per-test dollar costs…
23.9🔒Exercises[Beginner] Exercise 23.1: Walk your current project through the decision tree in 23.1. Does the recommendation match what you use? If not…
23.10🔒Career Translation- Developed a tool-selection decision framework for AI-driven browser automation, evaluating Playwright CLI/MCP, Stagehand, browser-use…
23.11🔒Q&AInterview Depth CheckPrompt: A team runs 500 Selenium tests nightly and wants to modernize. Decide between the Playwright CLI path, MCP, or a hybrid.