• 53% of surveyed developers now have AI agents in production or broad use, led by Thailand (64%),  Vietnam (59%) and Singapore (58%) but only 38% consider their codebase ready for full autonomy.  • Cost is now the top barrier to broader AI adoption, cited by 28% of developers as the primary  obstacle to wider AI agent use.  

• The share of developers saving seven hours or more a week across Asia has tripled to 55% from 18%  in 2025. 

• 71% expect AI agents to handle most development work within three years, though 79% still require  human approval for production deployment.  

• Thailand reported the region’s largest productivity rates, with 70% of developers saving at least  seven hours a week, followed by Vietnam (63%) and the Philippines (57%).  

Singapore, September 25, 2026 — Digital travel platform Agoda has released a new study showing that  software developers across Southeast Asia and India are rapidly transitioning from using AI assistants  for individual tasks to deploying autonomous AI agents across the software-development lifecycle. 

The Agoda AI Developer Report 2026 explores how this transition is reshaping the way developers work,  how organizations manage costs, and where human accountability sits. 

Drawing on extensive input from the developer community across Indonesia, Malaysia, Thailand, the  Philippines, Singapore, Vietnam, and India, the report examines the opportunities and challenges shaping  AI’s next phase across these markets.  

AI is delivering more value, but developers still verify the output 

Developers are experiencing greater productivity gains from AI. More than half (55%) now save at least  seven hours each week, compared with 18% in 2025. Gains were strongest in Thailand (70%), Vietnam  (63%) and the Philippines (57%). Reported code usability has also improved, with 62% saying AI-generated  code is usable often or almost always without major changes.  

AI is also being used more broadly across development work, including code review, architecture,  debugging, learning, and documentation. Yet human judgment remains essential: 86% of respondents still  review or validate AI-generated outputs always or most of the time.  

Agents are arriving fast, but cost is holding back wider adoption 

More than half (53%) of respondents say AI agents are already used in selected production workflows or  broadly across their organization. Thailand leads at 64%, followed by Vietnam (59%) and Singapore (58%).  However, only 38% describe their codebase as mostly or fully ready for a fully autonomous agent.  

Cost is the primary barrier to wider adoption, cited by 28% of respondents, ahead of integration  complexity (24%) and lack of governance (19%). The report finds that AI costs extend beyond model calls  to include the work needed to design, integrate, review, and validate agents. Four in five respondents said  they are currently working under some degree of AI usage limits, such as token quotas, usage caps, or  budget restrictions. 

Autonomy is growing, while accountability remains human 

Developers are open to granting AI agents more autonomy for lower-risk, reversible work such as  documentation. But as work moves closer to production, the demand for human oversight rises: 79%  report requiring human approval for production deployment, rising to 84% in both Vietnam and Thailand. 

AI is also changing the developer’s role. System architecture and design, AI literacy, and agent  orchestration are emerging as future skills. The transition is uneven, however: 49% of junior developers  report feeling less secure about their future career prospects, compared with 16.7% of CTO and VP  Engineering respondents. 

Idan Zalzberg, Chief Technology Officer at Agoda, shared, “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. 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 organizations 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.”  

The study was commissioned by Agoda and conducted in partnership with Macramé Consulting. As a  digital travel and technology company in the region, Agoda is committed to supporting local technology  talent, fostering innovation, and sharing insights that help developers and organizations build responsible  AI practices.  

Agoda’s commitment to innovation extends beyond AI, offering over 6 million holiday properties, more  than 130,000 flight routes, and over 300,000 activities, all of which can be combined in the same booking.  

For more information and to download the free report, visit