The discussion was under the ATTS digital transformation workstream, an ASEAN-ISIS initiative coordinated by the Singapore Institute of International Affairs (SIIA) and supported by the ASEAN Secretariat, which TFGI leads. Academics, practitioners and private-sector representatives took part, representing Singapore, the Philippines, Thailand, Viet Nam, Lao PDR and Timor-Leste. The stakes are considerable. Micro, small, and medium-sized enterprises (MSMEs) account for 85% of employment and 45% of GDP across ASEAN, and AI adoption could contribute up to 18% to the region’s GDP if the transition is inclusive. Readiness across the region remains uneven, and no widely used index directly measures SME-level readiness, leaving governments without a common basis for deciding where support should go.
Three priorities emerged for the region to take forward:
- Shifting support from individual task-saving tools towards the transformation of whole business processes;
- Pairing readiness assessment with tailored, sector-specific interventions rather than generic curricula; and
- Scaling responsible-AI obligations to the risk of the use case rather than the size of the firm
Together, these point towards an AI skilling agenda that meets SMEs where they are and stays with them as they move.
Participants and Facilitators
Participants
- Kirthi Pathania, Regional Director APAC, InspireXT, Singapore
- Jake Ervin Jonathan Go, Managing Partner, Springboard Philippines; Educator, De La Salle University, Philippines
- Ophakorn “Jim” Kouphokham, Founder and Managing Director, XM Technovator, Laos
- Ramon Cabrera, Faculty Member, University of Asia and the Pacific, Philippines
- Laurence Liew, Director of AI Innovation, AI Singapore
- Dr Dinh Khanh Le, Researcher, IPSS, Viet Nam
- Piyawan Chaiwisessakul, Senior Manager, Kenan Foundation Asia, Thailand
- Adilsonio da Costa Junior, Researcher, MDI-SIS, Timor-Leste
- Gayathri Haridas, Head of Future of Work, Access Partnership, Singapore
- Prof Matthias Tietz, University of St.Gallen, SGI-HSG Singapore
Convenor and facilitators
- Arifah Sharifuddin, Institute Director, Tech for Good Institute
- Fairoz Ahmad, Programme Fellow, Tech for Good Institute
- Erlanggasakti Ubaszti Putra, Programme Analyst, Tech for Good Institute
- Basilio Claudio, Programme Associate, Tech for Good Institute
- Ila Jeanne Perez, Programme Officer, Tech for Good Institute
Key Takeaways
1. AI Creates Value by Reshaping Workflows, Not Micro-Tasks
Most SMEs use AI to save time on single tasks such as marketing, design and branding, but this does not create a long-term business advantage. Real financial return occurs only when AI transforms entire operations, such as reducing a four-day process to 30 minutes. This is complicated by the fact that SME owners experience shifts in their priorities throughout the year, which slows structural change.
Improving AI adoption requires two main steps. First, leadership training programmes such as Singapore’s Digital Enterprise Blueprint show that business owners need direct training to transform their operations. Second, research shared at the FGD also found that projects run with outside partners are about twice as likely to lead to adoption compared to those run alone.
2. Readiness Assessment First, Then Sector-Specific Support
Participants agreed that the most effective initiatives begin with assessing a company’s readiness. A small restaurant, a manufacturer, and a technology company should not be offered the same AI course, and a readiness assessment followed by differentiated support was the model most often cited as leading to greater success. Vietnam’s SME digital transformation plan shows the scale this entails: at least 500,000 SMEs are to receive readiness assessments, consulting, and training by 2030, with at least 300,000 supported in adopting digital platforms and AI applications.
Effective AI adoption relies on context-aware capacity building, as generic training yields limited success given the unique operational landscapes of different industries. Organisations benefit most from structured guidance grounded in sector-specific best practices, actionable frameworks, and documented outcomes. Thailand applies this through OTOP (One Tambon One Product) and OSMEP (the Office of SME Promotion). Through its Business Development Service, the government co-funds the cost for MSMEs in the cultural and handicraft sectors seeking to integrate generative AI technologies into their operations.
3. Trust Grows When Rules Match the Risk of the Use Case
Low trust remains the strongest predictor of whether a firm remains at the pre-adoption stage. SMEs tend to ask how their business data is used in AI models and who is accountable when the output is wrong. The levers discussed included certification or a published list of trusted tools, simple guidance rather than complex regulation, and a safe environment to try AI tools before committing to them.
Participants are broadly aligned on the principle of proportionality, emphasising that governance obligations should be calibrated to the specific risk profile of an AI application rather than to enterprise size. In this context, the ASEAN Guide on AI Governance and Ethics is a good starting point. It is a voluntary, risk-tiered guideline designed to help MSMEs reduce compliance costs and work from a common regional baseline rather than fragmented national rules. Next steps to more effective operationalisation of the guide are simplified documentation templates and a more accessible sandbox mechanism for smaller businesses.
