Everything on Metrics & KPIs

Token counting is an ineffective metric for measuring AI engineering success. More valuable metrics include feature velocity, code acceptance rates, and the length of autonomous sessions, which better reflect innovation and efficiency.

Impact: Adopting these new metrics helps organizations focus on business outcomes and innovation speed rather than just cost optimization, leading to more strategic AI investments.

— from Roblox AI Native Engineering Strategy · Engineering Enablement by DX· Oct 30, 2026

Token consumption is no longer a valid metric for AI success. High usage often correlates with inefficiency rather than productivity, making it a directionless stat for business value.

Impact: Shifting to value-based metrics allows finance and engineering leaders to align AI spend with actual business outcomes, improving ROI justification.

— from AI Model Economics and Sustainable Engineering Practices · Dev Interrupted· Jun 12, 2026