AI’s Dual Impact on DevOps: Boosting Productivity While Slowing Delivery

Introduction

On October 22, 2024, Google released its annual DevOps research and assessment ( DORA ) which found that while generative artificial intelligence ( AI ) modestly improves productivity, it appears to also slow software delivery speed.

AI Handles at Least One Routine Task

The report discovered that more than three‑quarters of surveyed software developers rely on AI for at least one daily professional responsibility, with productivity gains of 7.5% in documentation volume, 3.4% in code quality, and 3.1% in code‑review speed.

AI‑driven workflow enhancements: 75.9% of developers use AI for routine tasks such as code writing, summarization, debugging, and testing. AI also speeds up code reviews and reduces technical debt, improving documentation quality by 7.5% when fully integrated.

However, the global survey also found a decline in delivery throughput of 1.5% and a drop in delivery stability of 7.2% .

Google Cloud’s DORA lead Nathan Harvey said the root causes of these declines are still unclear, possibly because:

AI challenges: While AI raises individual productivity, it is linked to the 1.5% reduction in throughput and 7.2% reduction in stability. Faster code generation leads engineers to create larger change sets in shorter time, increasing deployment complexity and failure risk.

Code generated by AI still requires engineer review and fixes before production deployment.

The report advises that organizations should not abandon AI but focus its investment on areas where clear benefits are evident, ensuring any AI‑generated code is carefully reviewed and tested.

The latest DevOps research indicates that applying AI in DevOps is closely tied to three pillars: experimental culture, continuous learning, and redefining how software is built.

Platform Engineering Rises but Has Untapped Potential

Platform engineering has become a key discipline for high‑performance teams. The 2024 DORA report highlights its advantages and complexities. Internal developer platforms (IDPs) help teams streamline workflows and scale self‑service capabilities.

Productivity gains: Teams using IDPs improve team performance by 10% and individual productivity by 8% . By abstracting complex operational tasks and enabling self‑service, platform engineering lets developers focus on coding and innovation rather than infrastructure management.

Although developer efficiency rises, performance appears to decline in some aspects.

Unexpected drawbacks: Platform engineering introduces efficiency challenges.

The DORA four metrics—lead time for changes, deployment frequency, change failure rate, and mean time to restore—are all lagging indicators.

The report found that when teams are required to use internal platforms for all lifecycle tasks, throughput drops by 8% and change stability falls by 14% .

This suggests that while platforms boost productivity, they can also increase pipeline complexity, slowing deployment speed and raising failure rates.

Overall, tracking these metrics helps DevOps teams reduce burnout and increase productivity (elite teams deliver software up to 182× faster than low‑performing teams), but each organization should consider them within its own business‑goal context.

Organizational Stability and Developer Burnout

The 2024 DORA report also emphasizes the importance of stable organizational priorities for maintaining developer productivity and well‑being.

Teams with relatively stable priorities outperform those with constantly shifting priorities, especially in reducing burnout.

Burnout and priority volatility: Teams with unstable priorities experience a 40% higher risk of professional burnout compared to teams with stable priorities.

Even with strong leadership and good internal documentation, such instability leads to a noticeable drop in productivity.

Conclusion

Although early iterations of AI and platform engineering may not immediately produce large effects, AI’s potential is significant, and no DevOps team can afford to ignore it.

References

[1] DevOps research report: https://devops.com/survey-sees-steady-adoption-of-ai-among-devops-teams/