
Measuring AI impact: the DX framework

Measuring the impact of AI on software engineering – with Laura Tacho
Topic 3 of 14
Published Jul 23, 2025 · Clip 10:30–18:08 (7:38)
Laura explains the most common questions from engineering leaders: 'what should I be doing?' (AI is frontier work with no pattern-matching precedent) and 'how do I measure and prove my AI decisions?' She introduces the DX AI measurement framework built with Abby, organized around utilization, impact, and cost, drawn from work with hundreds of companies. She criticizes output metrics like lines of code and acceptance rate, arguing quantity of code is disconnected from business impact and that source code is a liability. Acceptance rate only signals whether tools are fit for purpose, not whether they create velocity or innovation gains.