AI-assisted software development in 2025: Inside this year's DORA report
96 snips
Dec 11, 2025 Chris Westerhold, Global Practice Director for Engineering Excellence at Thoughtworks, dives into the transformative landscape of AI-assisted software development by 2025. He highlights how AI’s pervasiveness can either enhance or degrade quality if not measured carefully. Westerhold underscores the need for test-driven development and cautions about AI amplifying current practices, good or bad. He discusses the impact on junior developers, the importance of user-centric platforms, and the delicate balance of costs in adopting AI technologies.
AI Snips
Chapters
Transcript
Episode notes
AI Ubiquity Means Usage Quality Matters
- AI is ubiquitous in software engineering; the question is how well you use it.
- Measure impact with DORA metrics and supplementary measures for a holistic view.
Prioritize Quality Over Raw Throughput
- Avoid increasing throughput at the cost of product quality.
- Focus on meaningful tests, not just higher test counts or vanity metrics.
Drive AI With Developer Thoughtfulness
- Use TDD and keep engineers driving design even when AI generates code.
- Verify AI output to avoid fallbacks that mask real problems.
