AI as an Amplifier in Software Development
AI improves software delivery only when teams have the right systems and culture. The 2025 DORA report shows that success with AI depends less on tools and more on disciplined engineering, data quality, and platform strength.
AI use in software development is nearly universal. The 2025 DORA State of AI-Assisted Software Development report found that 90% of developers use AI in their daily work, with most reporting productivity gains. Yet many teams still face instability and uneven results. The reason is structural: AI amplifies whatever is already present. It strengthens strong systems and exposes weak ones.
The report identifies seven key practices that distinguish organizations that benefit from AI from those that struggle.
1. Establish a Clear AI Policy
Define how AI should be used, which tools are approved, and what review processes apply. Clarity prevents friction and builds trust in AI-assisted work.
2. Invest in a Strong Internal Platform
Ninety percent of surveyed organizations now use platform engineering. Teams that treat their internal platform as a product, focused on reliability, usability, and automation, see greater returns from AI. A well-built platform reduces fragmentation and ensures consistent workflows.
3. Maintain a Healthy Data Ecosystem
AI systems depend on accessible, accurate, and well-structured data. Unifying documentation, telemetry, and code repositories gives AI context and improves the quality of generated output.
4. Deliver in Small, Measurable Batches
Smaller changes lower risk and create faster feedback loops. When AI increases code throughput, this discipline prevents instability and allows teams to learn from each release.
5. Use Value Stream Management
Value Stream Management (VSM) connects AI-driven productivity to real outcomes. It ensures that local improvements in speed or automation translate into better team performance, user satisfaction, and product reliability.
6. Balance Productivity with Learning
The DORA report notes that while AI increases output, it can reduce hands-on learning. High-performing teams measure skill development and encourage critical engagement with AI suggestions rather than blind acceptance.
7. Keep Work User-Centered
AI can accelerate poor priorities as easily as good ones. Teams that maintain a clear understanding of user needs see higher quality outcomes and lower rework rates.
The DORA researchers emphasize that AI success is a systems problem, not a tools problem. High-quality platforms, clear governance, and continuous improvement practices amplify the benefits of AI while containing its risks. With these foundations, AI becomes not just a faster way to code but a multiplier for organizational performance.
Reference:
2025 DORA State of AI-Assisted Software Development Report, Google Cloud, 2025 — dora.dev