AI speeds coding but costs Singapore teams a day weekly
Software development teams in Singapore are gaining speed from artificial intelligence, but are losing a full working day each week to inefficient workflows and fragmented tools, according to new research from GitLab.
The 2025 DevSecOps report surveyed 3,266 IT operations, security and software development professionals worldwide. The sample included 252 respondents based in Singapore.
GitLab describes the current phase as an "AI paradox" for DevSecOps teams. Coding is faster, but operational complexity is rising.
The study found that Singapore practitioners lose an average of eight hours every week due to inefficient processes. This is one hour more than the global average.
Respondents highlighted several main sources of lost time. They cited a lack of knowledge sharing, a lack of cross-functional communication, and inefficient or unclear work processes.
Each of these factors ranked above 30% in Singapore. Lack of knowledge sharing and lack of cross-functional communication were both reported by 34% of respondents. Inefficient or unclear processes were reported by 33%.
Tool sprawl is another feature of the local environment. The report said 62% of Singapore respondents use more than five tools for software development.
Use of AI tools is equally fragmented. It found that 54% of Singapore respondents use more than five AI tools.
The research suggests a growing preference for platform-based approaches. It reported that 82% of Singapore DevSecOps professionals agree that agentic AI will be most successful when implemented within a platform engineering model.
GitLab said AI is also reshaping roles and skills across software teams. Many respondents expect more engineers to enter the field as AI changes coding tasks.
In Singapore, 79% of professionals surveyed agree that there will be more engineers as coding gets easier with AI. The report stated that 87% believe software engineers who adopt AI are future-proofing their careers.
Respondents also anticipate major changes to their own jobs. The survey found that 81% think AI will significantly change their roles within the next five years.
Many professionals want more formal support with this shift. According to the research, 87% wish their organisations invested more in helping them upskill around AI and related tools.
Human oversight
AI adoption is now widespread along the software development lifecycle in Singapore. The report found that 98% are using or planning to use AI in software development work.
Respondents indicated a growing comfort with delegating tasks to AI. On average, DevSecOps professionals in Singapore would trust AI to handle 40% of their daily work tasks without human review.
However, most have also seen problems in AI-generated outputs. The report said 66% have experienced issues with code created through so-called "vibe coding".
The research also explored views on human qualities in software work. It found that 88% of respondents agree there are essential human qualities, including creativity and innovation, that agentic AI will never fully replace.
Rising compliance strain
The study reported that regulatory and compliance pressures are increasing. Many teams see AI as a factor that makes compliance work more complex.
Among Singapore respondents, 71% agree AI is making compliance management more challenging. At the same time, many organisations are identifying issues late in the lifecycle.
The research stated that 76% report more compliance issues are discovered after deployment than during development. This pattern is pushing interest in code-level governance approaches.
Respondents ranked security and compliance skills as central to their future careers. The report found that 40% believe implementing AI for security and compliance is the top skill required for advancement.
This ranked ahead of incident response and recovery skills. It also ranked ahead of fluency in programming languages.
Professionals expect a significant shift in how compliance is handled over the next few years. By 2027, 85% of respondents predict that compliance will be built into code and automatically applied.
GitLab said these findings show that AI is compressing development timelines while exposing weaknesses in existing toolchains and governance models.
"This survey illustrates what we call the 'AI Paradox,' where coding is faster than ever, yet the lack of quality, security, and speed across the software lifecycle is causing friction on the road to innovation," said Manav Khurana, chief product and marketing officer, GitLab. "Toolchain fragmentation has created bottlenecks for developers, and AI agents are amplifying the issue. Organisations need a new framework to match the speed of software development in the age of AI, one that provides intelligent orchestration across the entire software lifecycle while addressing the interconnected requirements of AI orchestration, governance, and compliance that individual point tools simply cannot solve."
