STEP 05 — RECOMMENDATIONS
Where AI creates capacity — and what to do with it
This isn't a headcount cut. It's a map of what each person is freed up from, and what they should do more of instead.
Summary
Across the 8 of 8 roles scored so far, AI is estimated to free up ~2.0 of 8.0 FTE (25%) of team capacity. The single biggest opportunity is Diego Alvarez (QA Engineer), where 40% of the role is automatable today. 3 reskilling priorities are outlined below to help the team capitalise on that freed-up time.
Not a reduction — reinvested into the higher-value work below. No one's role shrinks to a fraction of a job; the work each person does shifts.
Elevated roles
Alex Kim — Senior Software Engineer
AI now handles
First-pass code review and boilerplate generation.
More time for
System design, mentoring, and cross-team technical decisions.
Ben Foster — Product Manager
AI now handles
Status reports and competitive research summaries.
More time for
Competitive strategy and roadmap trade-off decisions.
Diego Alvarez — QA Engineer
AI now handles
Routine test execution and reporting.
More time for
Security and performance testing depth.
Grace Liu — Product Manager
AI now handles
Spec first-drafts and analytics summaries.
More time for
Customer discovery, stakeholder alignment, and prioritisation calls.
Jordan Lee — Software Engineer
AI now handles
Feature scaffolding and unit test generation.
More time for
Feature complexity, code review depth, and technical debt work.
Maya Chen — QA Engineer
AI now handles
Regression test generation and flaky test triage.
More time for
Exploratory testing, test strategy, and quality process design.
Nina Torres — Software Engineer
AI now handles
Boilerplate CI/CD scripting and routine debugging support.
More time for
Pairing, onboarding support, and edge-case handling.
Sam Patel — Software Engineer
AI now handles
UI scaffolding and component test generation.
More time for
Design system quality and complex interaction work.
Reskilling plan
AI-assisted development fluency
Using AI coding assistants effectively for scaffolding, first-draft code, and test generation — with strong review discipline.
Whole team
AI-output verification & review
Critically reviewing AI-generated code, tests, and specs rather than accepting them at face value.
Whole team
Judgement-first system thinking
Investing freed time in architecture, quality strategy, and product judgement that AI can't replace.
Whole team