DEVOPSdigest

While AI is rapidly reshaping roles in DevOps, most enterprises struggle with governance, according to new research from Enterprise Management Associates (EMA) and DEVOPSdigest.

62% of the IT leaders surveyed cited security and privacy risks as their top concerns, according to the report AI in DevOps: Adoption Outpaces Governance and Changes the Role of the Developer.

AI ROI Tracked in Engineering Teams with Emphasis on People and Process

Organizations primarily measure ROI through code quality and defect reduction (70%) and developer productivity (62%), with fewer tying AI benefits to release velocity or business outcomes — indicating ROI is still framed mostly in engineering terms rather than enterprise value. However, realized outcomes from the survey also indicated better people and processes through AI, including:

■ 62% higher developer satisfaction

■ 49% faster time to market

■ 43% faster onboarding with junior developers

■ 38% reduced toil in DevOps or pipeline tasks

Duality of Impact Reveals Governance Needs for AI in Continuous Software Quality

Despite widespread AI adoption, the research identifies substantial governance gaps. While AI development tools deliver measurable benefits, including improved test coverage (56%) and more consistent code formatting (55%), a range of concerns confirms the need for governance:

■ 52% report concerns about introducing security vulnerabilities, and 52% report concerns around code defects and bugs.

■ 69% of respondents were concerned about over-reliance on AI tools, while 61% expressed concerns about "blind faith" in AI results. 

■ Over half of the respondents (57%) report a negative or neutral experience with an AI tool hurting their workflow, primarily due to inconsistent results in code quality and testing. 

Vibe Coding Seen as Beneficial, But Quality and Control Risks Persist

51% of those surveyed actively use vibe coding for new features, and most see strong flow benefits — but nearly half stress the need for oversight by experienced developers to correct inefficiencies and avoid structural risks. While adoption remains strong with vibe coding, only 3% of those surveyed felt like it would reshape developer workflows.

Vibe coding's top benefits according to respondents

■ Enhancement of productivity (cited by 38%) 

■ Helpful bridge to lower barriers for new or returning developers (cited by 33%) 

■ Useful for business users and less relevant with developers (cited by 20%)

Control and Quality are the Top Challenges around Vibe Coding

■ Risk of poor quality or insecure code (54%)

■ Over-reliance on inexperienced developers (48%)

■ Lack of integration with existing DevOps/CI/CD pipelines (45%)

■ Limited control over generated architecture or design choices (44%)

■ Unclear IP ownership or generated applications (33%)

Enterprises' Investment Focuses on Closing Interoperability Gap, Addressing Tool Sprawl, and Orchestration Challenges and Opportunities

"Beyond security risks, the biggest barriers to AI in DevOps are tool sprawl, interoperability gaps, and the difficulty of orchestrating AI assistants across workflows," said Dan Twing, President and COO of EMA. "Without unified governance and vendor ecosystems that simplify integration, AI expansion will remain fragmented — limiting both scalability and return on investment."

Key Factors that lead to these conclusions include:

■ 48% "Strongly supportive" and 27% "Somewhat supportive" of using non-approved tools. Demonstrates decentralized experimentation and fragmented control leading to tool sprawl and lack of centralized control.

■ 45% cite security concerns, 32% IP protection, and 25% vendor conflict when it comes to why organizations are blocking AI usage, which highlight integration and interoperability barriers during production rollout.

Survey respondents listed the top improvements they want with their AI-powered development workflows

■ 55% indicated a need for real-time vulnerability detection

■ 53% wanted automated test generation

■ 46% enhanced DevOps pipeline orchestration

■ 46% wanted AI assisted performance testing or environment simulation 

Methodology: The research is based on a survey of 206 IT executives and technical leaders.