Singapore – Software developers across Southeast Asia and India are adopting AI agents across the software development lifecycle, although many organisations are still preparing their codebases and processes for greater autonomy, according to Agoda’s report.
The study, commissioned by digital travel platform Agoda and conducted in partnership with Macramé Consulting, draws on responses from developers in Indonesia, Malaysia, Thailand, the Philippines, Singapore, Vietnam and India. It examines how AI is changing development practices, the costs associated with adoption and the role of human oversight.
The report found that 53% of respondents were using AI agents in selected production workflows or broadly across their organisations. However, only 38% considered their codebase mostly or fully ready for a fully autonomous agent.
Productivity gains and continued human review
More than half of respondents (55%) said AI saves them at least seven hours a week, up from 18% in 2025. Meanwhile, 62% reported that AI-generated code was often or almost always usable without major changes.
Developers are using AI for a wider range of tasks, including code review, architecture, debugging, learning and documentation. Despite these gains, 86% said they always or most of the time review or validate AI-generated outputs.
Cost remains a barrier to agent adoption
Cost was the most frequently cited obstacle to broader AI agent use, identified by 28% of respondents. Integration complexity followed at 24%, while 19% cited a lack of governance.
The report notes that AI-related expenditure extends beyond model calls to include the work involved in designing, integrating, reviewing and validating agents. Four in five respondents said they were operating under some form of AI usage restriction, such as token quotas, usage caps or budget limits.
Human approval remains important
While developers expressed openness to giving agents more autonomy for lower-risk and reversible tasks, such as documentation, human oversight remained common for production-related work. Some 79% of respondents said human approval was required before production deployment.
The report also identifies system architecture and design, AI literacy and agent orchestration as skills likely to become increasingly important for developers.
However, confidence about future career prospects varied by seniority. Nearly half of junior developers (49%) said they felt less secure about their career prospects, compared with 17% of chief technology officers and vice-presidents of engineering.
“Artificial intelligence is reshaping how developers across Southeast Asia and India build, learn, and collaborate. What began as a way to speed up individual tasks is becoming a broader change in how software is built,” said Idan Zalzberg, Chief Technology Officer at Agoda.
“Across industries, developers are seeing stronger productivity gains and are beginning to use agents across more of the development lifecycle. But greater capability brings new responsibilities. The real cost is not simply the model call. It is the work required to make AI trustworthy in production. The organisations that will excel in this next phase will be the ones that manage cost carefully, build the right safeguards, and keep human judgment where it matters most,” Zalzberg continued.
Agoda said it supports local technology talent and innovation, and shares insights intended to help developers and organisations establish responsible AI practices.
