Adoption is not capability – the real AI test for Singapore’s SMEs

Singapore will assume the Asean chairmanship in 2027, and the bloc's 11 member states are expected to sign the Digital Economy Framework Agreement (Defa) in November, establishing common rules for digital trade and AI capability across the region. A study commissioned by Amazon Web Services (AWS) covering 1,500 Singapore firms found that three-quarters of financial-services SMEs use AI, but in 60% of firms, AI work stalls if one key person leaves. AWS has committed US$1 billion to embed forward deployed engineers in customer teams to ensure capability transfer.

Adoption is not capability – the real AI test for Singapore’s SMEs The Republic’s AI scoreboard looks strong. But, in many firms, the whole capability rests on one person SINGAPORE will assume the Asean chairmanship in 2027. But before that, however, the bloc’s 11 member states are expected in November to sign the Digital Economy Framework Agreement Defa , its first comprehensive, region-wide pact dedicated to the digital economy. Defa establishes common rules that facilitate digital trade, enable trusted cross-border data flows, and crucially, build enduring artificial intelligence capability across a region of more than 600 million people. But a framework carries only what its members can supply. Singapore has moved fast on AI adoption, and the numbers are genuinely strong. The next step is making sure that progress is deep enough to travel, and that is a harder, less-visible kind of work. The missing layer in AI diffusion Singapore is not short on ambition or support. The Economic Strategy Review signalled the importance of developing leading firms as ‘champions of AI’ and a market for a new class of “forward deployed engineers” has opened up in swift succession – companies are hiring people who combine software skills with the ability to explain technology to clients and solve everyday operational problems. In a study commissioned by Amazon Web Services AWS covering 1,500 Singapore firms, three-quarters of financial-services small and medium-sized enterprises SMEs say that they use AI, and so does a majority in healthcare and manufacturing. That momentum is an asset, not a given. The question is no longer whether firms adopt, but whether the capability runs deep enough to outlast any single champion. Our research found that in 60 per cent of firms, it does not yet: remove one person and the AI work stalls or stops entirely. The people who get the most from AI did not build it Earlier this year, a global AI hackathon drew thousands of entrants. The top finishers were not developers – they were a cardiologist and a lawyer, domain experts who had never shipped production code before. They won because they understood problems their industries had lived with for years, that knowledge, not syntax, was the multiplier. Case in point: the capability that matters most rarely sits with the engineers. The people who build AI models are seldom the ones who get the most from them. The value lands with the people who know the business: the plant supervisor who knows why a schedule slips, the relationship manager who understands a borrower better than any model. When that person is not in the room, a firm loses more than a pair of hands. It loses its translator, the one who can judge whether what the AI produced makes sense for this customer, this shipment, this patient. On its part, AWS recently committed US$1 billion to embed forward deployed engineers directly inside customer teams, not to build systems for them but to leave them self-sufficient when the engagement ends. The bet reflects a broader understanding: tools without capability transfer do not stick. While the national conversation is mostly about growing tech talent, the better question for most SMEs is whether the people who already understand the business are using these tools. That shift, from one person’s know-how to a team’s habit, is what turns adoption into capability. Capability is not built in a classroom Here, our instincts can work against us. Faced with a skills gap, we train: courses, playbooks, showcases of firms that got it right. All useful, but limited. In the Singapore Business Federation’s National Business Survey 2025 , about a third of firms still doubted the accuracy and reliability of AI outputs, and that doubt is the heart of the matter. Judgment with these tools is not taught in a seminar. Watching a polished demo does not prepare you for making that decision, but running into the problem yourself does. A different kind of peer network So what would help? Government programmes have lowered the cost of tools, compute and expertise, and technology partners bring the infrastructure. What is missing is the layer in between: a place where firms exchange what they actually learned, including what failed and what cost more than expected. Singapore runs peer networks well, but they tend to showcase success, and only work if the right people are in the room. The programmes that produce lasting capability share one feature: they start with the person closest to the problem, not necessarily the person closest to the budget. That principle scales. Sector-based deep dives and hackathons for Asean’s SMEs, built around the problems frontline staff actually face, can spread fluency to the people using these tools rather than the lone champion who owns them. And Defa can be the umbrella to turn these events into a regional framework to build capability, not just adopt tools. Engineers can be deployed and networks can be convened. Now, Singapore needs to create the environment that produces more of these people, not by training everyone to code, but by giving domain experts the tools, the sandboxed space, and the institutional permission to build. User-level engagement is not just good HR practice. It is a precondition for regional scalability under Defa. Multiply them, document what they build, and Defa becomes the mechanism that carries proven capability from Singapore into the region. The writer is head of trade policy Asia-Pacific and head of public policy Singapore at AWS Decoding Asia newsletter: your guide to navigating Asia in a new global order. Sign up here to get Decoding Asia newsletter. Delivered to your inbox. Free. Share with us your feedback on BT's products and services TRENDING NOW /pulse?ref=trending-now Kallang Distripark set to be redeveloped for thousands of new private homes Financial planners urge investors to consider other options as T-bills yields lose steam Car-rental firm Charlotte Vehicle allegedly owes about S$140,000 to 34 customers Too little, too late? Manila’s billion-dollar bid to ignite its sputtering EV industry