1 Step 1: Set Up Your AI-Assisted Workflow Foundation You’ll learn how to structure prompts, context, and guardrails so AI outputs are consistent, reviewable, and easy to integrate into your normal dev flow. Do this exercise: Create a “prompt harness” template (Goal → Constraints → Inputs → Output Format → Checks) and use it on one real task this week.
2 Step 2: Requirements to Testable Stories (Fast, Not Sloppy) Backlog You’ll learn how to turn fuzzy ideas into crisp user stories with acceptance criteria that drive implementation and reduce rework. Do this exercise: Take one “confusing” backlog item and rewrite it with Given/When/Then acceptance criteria plus 3 edge cases.
3 Step 3: TDD with AI (Keeping You in the Driver’s Seat) TDD You’ll learn how to use AI to draft unit tests, generate cases, and refactor safely—while preserving intent and design quality. Do this exercise: Pick a small function, ask AI for test cases, then write the tests yourself and compare gaps.
4 Step 4: Design & Architecture Checks (Before You Code) Design You’ll learn lightweight architecture prompts that expose coupling, risks, and missing decisions before they become expensive. Do this exercise: Run an “architecture risk scan” prompt on a current feature and identify 2 decisions to document.
5 Step 5: Code Review with AI (Quality, Security, and Style) Quality You’ll learn how to use AI as a second set of eyes for readability, security hotspots, performance pitfalls, and missed tests—without rubber-stamping. Do this exercise: Paste a PR diff (or key files) and ask AI for “top 5 risks + suggested tests,” then validate each item.
6 Step 6: Repeatable Prompt Patterns for Your Team Scale You’ll learn how to standardize prompts into reusable patterns so the whole team can produce consistent results across stories and sprints. Do this exercise: Create a shared “Prompt Playbook” page with 3 prompts: story refinement, test ideas, and review checklist.
7 Step 7: End-to-End Delivery (Docs, Release Notes, Runbooks) Delivery You’ll learn how to use AI to produce the operational “last mile” artifacts—without losing accuracy or creating documentation debt. Do this exercise: Generate release notes from a changelog/PR list, then edit it to match your audience and verify claims.