
Audit one legacy feature to identify where AI can genuinely multiply your output
You'll have a written assessment of one shipping feature showing where AI accelerated work (boilerplate, test writing, refactoring) versus where human judgment was irreplaceable, plus time savings dat
What you'll make
AI Integration Assessment
- Feature Context: what you audited and why
- Acceleration Zones: where AI genuinely multiplied output
- Human Oversight Boundaries: where you had to intervene or validate
- Time Savings Analysis: concrete before/after measurements
- Replication Framework: your approach for evaluating the next feature
Write each section with specific examples from your actual feature audit and measurable time comparisons
The 3 parts
Each part ends with a check you can see.
Map Your Feature's Task Boundaries
Pick one shipping feature and break it down into discrete development tasks to identify where AI might accelerate versus where human judgment drives decisions.
Done when
A task breakdown of one feature with at least 8 named tasks categorized as 'boilerplate generation', 'logic design', or 'integration testing', with time estimates for each.
Test AI on Three Real Tasks
Take three tasks from your breakdown and actually use AI to attempt them, measuring time savings and quality compared to your manual approach.
Done when
Completion data for 3 tasks showing time before/after AI assistance, quality assessment of AI output, and specific edits you made to reach shipping standard.
Document Your Reusable Framework
Synthesize your findings into a written assessment and create a framework you can apply to evaluate AI integration opportunities in other features.
Done when
A complete AI Integration Assessment document with specific examples, time measurements, and a step-by-step framework for evaluating the next feature.
Capstone: you put together your AI Integration Assessment.
Where this helps
An elective: a single lesson that sits outside rtwo's job certificates and doesn't count toward one.
About this lesson
What you'll be able to do
- Produce a written assessment of one production feature showing where AI genuinely multiplied output versus where human judgment was irreplaceable
- Document specific time savings with concrete before/after comparisons
- Have a reusable framework for evaluating AI integration opportunities in existing codebases
Who it's for
- Developers working on legacy systems who want to understand where AI fits without rewriting everything
- Engineering teams evaluating AI tooling ROI on real shipping features
- Tech leads deciding which parts of their codebase to augment first
What it won't cover
- Advanced AI system architecture or agent framework design
- Comprehensive code architecture redesign or greenfield AI development
- Theoretical AI capability predictions or vendor feature comparisons
Skills you'll practise
- Integration impact measurement
- AI-human boundary identification
- Legacy code pattern recognition
- Time savings quantification