Use AI to improve delivery speed and quality, with human accountability

What this means in practice

Teams use AI to speed up thinking and delivery, but humans remain accountable for decisions, code quality, and outcomes. AI is used deliberately to increase clarity, quality, and learning. We use it where it reduces waste, improves correctness, or strengthens feedback loops.

Why this matters

AI can increase throughput and quality, but only if it is used with clear responsibility and good judgement. This principle prevents teams from treating AI outputs as truth, and keeps ownership, safety, and professionalism intact.

Practices that meet this principle

Generating test cases from acceptance criteria, then reviewing and refining before use
Reviewing acceptance criteria for ambiguity, missing scenarios, and edge cases
Drafting code with AI, then verifying behaviour, security, and performance
Using AI to support code review, but keeping final review decisions with the team
Using AI to accelerate defect fixing and refactoring, backed by tests and safe rollout
Documenting key decisions and trade-offs when AI materially influenced the outcome

Validation

A project meets this principle when:

A named owner remains accountable for each decision and change
AI-generated outputs are reviewed and validated before merging or release
Automated tests and/or other checks provide evidence of correctness
The team can explain the rationale and trade-offs for material decisions