“Developers are advancing AI adoption with peer oversight and disciplined practices, but the next step is to pair that momentum with enhanced confidence in its capabilities,” said Idan Zalzberg, Chief Technology Officer at Agoda.
INCONSISTENT OUTPUT REMAINS the primary concern on the adoption of artificial intelligence (AI) among developers in Southeast Asia, resulting in AI not getting full confidence on its usage.
This is among the findings of Agoda’s AI Developer Report 2025, which showed that 79 percent of developers citing unreliable results as the biggest barrier to deeper AI integration, outweighing concerns related to access, cost, or tooling.
“Developers are advancing AI adoption with peer oversight and disciplined practices, but the next step is to pair that momentum with enhanced confidence in its capabilities,” said Idan Zalzberg, Chief Technology Officer at Agoda.
This challenge is most acute in markets such as the Philippines (88 percent) and Thailand (84 percent), where concerns about output reliability are highest.
Even in more mature AI markets like Singapore (77 percent) and Malaysia (73percent), inconsistent results continue to temper confidence and reinforce the need for human oversight.
QUALITY CONTROL
In response, developers have adapted their workflows to maintain quality. Two-thirds report they always review AI-generated code before merging, and many routinely rework outputs until they meet production standards.
Instead of reducing accountability, the Report finds that AI use has increased the emphasis on review and human oversight. Verification and ownership have become more prominent in daily development work, with engineers retaining responsibility for final outcomes. Confidence, for now, is conditional, built through repetition, testing, and experience rather than assumed by default.
The findings suggest that AI maturity is no longer defined by adoption alone. While productivity gains are tangible, confidence grows only when outputs are consistent and predictable. Teams with strong review habits report higher confidence, while those without structured verification remain more cautious.
The Agoda AI Developer Report 2025 draws on survey responses from developers across Southeast Asia and India and includes insights from regional companies such as Carousell, MoMo, Omise, and SCB 10X. It provides a detailed view of how developers are integrating AI into their workflows, where confidence breaks down, and what it will take to move from widespread use to trusted maturity.
ADOPTION CONTINUES
The report said that developers across Southeast Asia and India are using AI at scale, but confidence in its reliability has yet to fully mature.
While AI has become a regular part of software development workflows, many engineers remain cautious, treating it as an accelerator rather than a dependable substitute for human judgment. AI adoption is now widespread across the region.
Nearly nine in ten developers say they use AI on a weekly basis, and most report clear productivity gains. However, confidence has not kept pace with usage.
“The next differentiator will not be who adopts AI first, but who builds a clear framework around it for consistent and productive usage,” said Zalzberg.
 among developers in Southeast Asia, resulting){kind=link}