RACER Framework
RACER is the guiding model for AI adoption in software engineering. It was developed by Plandek and adapted by SOK to reflect our specific context and goals.
Dimensions
Rollout
How widely AI tools are deployed and actively used across teams. Tracks adoption breadth — from individual use to team-wide and org-wide rollout.
Approach
The quality of how teams adopt AI. Looks at whether teams are using AI intentionally, sharing learnings, and building repeatable practices.
Constraints
The organizational, technical, process, security, compliance, and cultural barriers that block effective AI adoption. A strong constraints posture means barriers are visible, prioritised, owned, and actively resolved — not just documented.
Engineering Impact
The measurable effect of AI adoption on engineering work — velocity, quality, review time, test coverage, and other engineering metrics.
Results
The broader outcomes: business value delivered, developer satisfaction, and progress toward the program's strategic goals.