AI Assisted Engineering
summary
A 24-week adoption experiment with 100 people across 8 teams builds an EOM-based chat platform, Inject Alita integration, GitHub Copilot and an adoption dashboard plus three AI coaches to measure impact. Results show development time −6% testing time −16% testing queue +103% requirement prep −16.5% lead time +25%: coding and test generation speed up but handoff queues explode and overall lead time can increase. Conclusion: tools help, but process visibility and queue management determine whether team throughput actually improves.
AI-written summary · about
chapters (11)
- 0:00Introduction and welcome
- 4:44AI’s emotional impact and future
- 10:08Initial reactions to GPT models
- 13:35Engineering productivity concerns
- 21:05DORA research contradictions
- 26:19Three-month AI adoption experiment
- 32:57Generative AI adoption challenges
- 35:06Productivity gains and queue times
- 41:33Future of AI in software development
- 47:38Rollout and industry consensus
- 51:16Q&A on AI risks and new approaches
related talks
27mAI for software development: A reality check
16mEngineering Intelligence & Developer Productivity in the AI Era
29mLeadership in AI-assisted engineering
27mFrom AI Assistants to AI Coworkers: How Engineering Teams Ship Faster with Enterprise Context
32mAI and Engineering Closing the Gap Between Expectations and Reality 1
29mAI Conversation
24mDevelopers and AI Enemy or Frenemy Leveraging AI and Atlassian to empower developers
29mThe AI Teammate Is Here — Now What?