The Question AI Providers Hope VPs Of Engineering Never Ask
The Next Web, Monday, April 20th, 2026
AI coding adoption lacks outcome measurement, creating a costly blind spot as providers profit from token usage rather than production value.
Most engineering leaders measure AI coding adoption by usage metrics rather than actual production outcomes, creating a structural misalignment with AI providers who are incentivized to maximize token consumption regardless of code quality.
The article draws parallels to early cloud computing's FinOps evolution, warning that companies spending $86-$28,000 per developer monthly on AI tools have no visibility into what code actually reaches production. Engineering leaders need commit-level attribution and measurement systems that connect AI spending to real business impact, allowing them to optimize costs and identify which vendors produce usable code versus create rework burden.