REDPIXEL - stock.adobe.com

Most developers in Southeast Asia and India use AI tools

A new study finds 95% of developers in Southeast Asia and India use AI tools, but many of their employers lack formal AI policies, leaving them to forge their own paths in upskilling and governance

While nearly all developers in Southeast Asia and India are using artificial intelligence (AI) tools in their work, they continue to face challenges associated with the use of the technology, a new study by digital travel platform Agoda has found.

According to Agoda’s AI developer report 2025, 95% of developers surveyed use AI every week, driven primarily by the need to speed up software development, with 37% of engineers saving four to six hours per week.

However, only 22% use AI tools to solve unfamiliar problems, and just 43% believe the tools can perform at the level of a mid-level engineer. And while 94% use AI to generate code, usage drops for downstream tasks such as documentation, testing, and deployment.

Such behaviour has led to a culture of “pragmatic optimism”, the report noted, where developers use AI as a productivity tool but rely on human oversight to ensure quality. Indeed, 67% of developers review all AI-generated code before merging, and 70% routinely correct outputs.

That developers are reviewing AI outputs is critical, given that about 60% of organisations have no formal AI policy. Nearly three in four (72%) developers reported productivity gains and better code outcomes from team-led reviews and validation processes, proving that human oversight remains key to responsible AI adoption.

Despite the keen adoption of AI among developers, the report identified a growing skills gap, with training struggling to keep pace. Most developers (72%) are self-taught, picking up AI skills through online tutorials and personal projects. Only 28% receive formal, employer-led AI training.

The report noted that access to structured training varies by market. In Singapore, developers are almost twice more likely than those in Vietnam to have access to formal AI training programmes. The result is a workforce that’s not only learning faster than organisations can teach, but also self-propelled and increasingly independent, it added.

Developers are also under pressure to keep up. In Singapore, 44% of developers worry about falling behind if they don’t adapt to AI’s pace of change, and 58% of all respondents now believe AI proficiency should be a baseline hiring requirement.

“AI is reshaping how developers across Southeast Asia and India build, learn, and collaborate,” said Idan Zalzberg, chief technology officer at Agoda. “What began as a way to speed up tasks like writing, testing, or debugging code has grown into a broader shift in how software is built. Today, AI helps teams move faster, learn continuously, and solve problems in new ways.

“In this region, AI use is mainstream but still uneven. Developers are approaching AI with pragmatism – accelerating work, maintaining quality, and experimenting thoughtfully rather than replacing skill or judgement. The real opportunity lies in supporting this ground-up maturity with structured practices and responsible experimentation, turning high adoption into consistent, lasting capability,” he added.

How developers are using AI

The report singled out several companies across the region that are embedding AI into their workflows to boost productivity while maintaining high standards.

At Singapore-headquartered marketplace Carousell, engineers are saving four to five hours per week using AI. The technology is used to automatically categorise products from images, detect spam listings, and identify potential scams in user chats, among other use cases.

Rajath Ramesh, Carousell’s group director for product and platform engineering, noted that the biggest benefit of AI is not just speed. “AI gives us back hours we can use for design, problem-solving, and mentoring,” he said.

Agoda itself has developed an internal AI tool dubbed CodeBuddy to help with code reviews and catch potential bugs before they reach users. The tool has reduced the volume of human comments needed per merge request by over 30%, speeding up deployment time.

In Vietnam, super-app MoMo uses a tool called Figma2Code, which converts design files directly into mobile application code, cutting component review time from up to three hours down to just 30 minutes.

Meanwhile, SCB 10X, the innovation arm of Thailand’s Siam Commercial Bank, uses its own Thai large language model (LLM), Typhoon, alongside Copilot, ChatGPT, and Gemini, to assist developers with code generation, debugging and quality control, saving them five to six hours weekly.

Read more about AI in APAC

Read more on Software development tools