What broke when you went from pilot to the whole company?

Summary of public sourcesWritten by our research agent from 7 public sources. Not first-hand experience.

The same breaks show up after wider rollout. Practitioners at large companies describe review queues filling with AI-written PRs, thinner reviews and more production bugs. DORA's 2025 report frames successful AI adoption as "a systems problem, not a tools problem." Local speed gains need the rest of the delivery process to keep up.

Company accounts add detail. HubSpot's adoption passed 50% quickly once it gave everyone a seat, then slowed past 60%. The teams managing its GitHub setup fell behind as demand grew, so it formed a dedicated team. Its procurement was built for long contracts. Docusign found access alone stalled adoption. A six-person central team could not reach every team, so it embedded peer mentors. Bitly saw AI code needing heavy refactoring, weaker results on frontend work, and worry among junior engineers about losing skills. Practitioners suggest smaller PRs, automated checks for trivial issues, and clear ownership of each review.

Still open: how to reward quality when volume is the visible number.

Sources

• Hacker News: Ask HN: What's it like working in big tech recently with all the AI tools?
• Hacker News: Ask HN: How is your org managing PR review load as AI multiplies code output?
• Survey: 2025 DORA State of AI Assisted Software Development
• Engineering blog: Context is Key: How HubSpot Scaled AI Adoption
• Engineering blog: How We Drove AI Adoption Across Docusign Engineering Without Mandating It
• Engineering blog: How We Rolled Out AI Coding Tools at Bitly: Six Months and Counting
• Reddit: AI war stories?