Structured rollouts and Indeed case study

Measuring the impact of AI on software engineering – with Laura Tacho

Topic 12 of 14

Published Jul 23, 2025 · Clip 58:52–1:07:27 (8:35)

Laura observes that highly regulated industries (finance, insurance, pharma) get the best AI rollout results because they must be deliberate and structured—acceptable use policies, licensing, budgeting, and PII protections. She highlights Indeed's structured, experimental approach: cohort trials of multiple tools with comparative analysis, hypothesis-driven experiments like using AI code review to shorten high-latency feedback loops across time zones, and identifying high-impact use cases such as unit tests and migrations. The host and guest discuss AI-assisted migrations (including AWS Q Developer's Java upgrades) as ideal 'undifferentiated heavy lifting,' and Laura shares a tip: do one migration by hand, then ask the model to generate a prompt reproducing that diff for subsequent files.