Paving the SME Roadmap for AI Transformation at AI Festival Asia 2025

With the deluge of AI tools trumpeting to change the game today, many businesses could be caught in a loop of trying to crack the code to sustainable implementation of AI. 

In Singapore, this means the majority of Small & Medium Enterprises (SMEs), a group that comprises 99% of all businesses in the country. Despite most decision-makers believing AI to be critical in reaching strategic goals and sharpening their competitive edge, Singapore companies face similar challenges with the rest of the world in matters of Data Governance, budget and having quality data for AI projects. 

For a topic so ostensibly well-wrung in both its technological promise and operational roadblocks, is there more to be said about the path to adopting AI in a well-governed and resource-effective way? Definitely.

It was with this premise in mind when we accepted an honoured invitation by the Association of Small & Medium Enterprises (ASME) to speak at the inaugural AI Festival Asia (AIFA) 2025, organised in conjunction with the Lifelong Learning Institute (LLI). A valued partner of Straits Interactive, ASME shares in our mission to elevate SMEs with AI capabilities to meet the dynamic challenges of the business landscape. This year's theme, "Evolve, Digitalise, Secure," captured the pressing need for businesses to embrace the AI revolution in tandem with responsible AI practices and privacy guardrails.

Across the two-day event, we contributed to the conversation through panel sessions with other experts in AI technology and Data Governance, and leaders advocating for the digital transformation of SMEs in Singapore. 

The Six Cs of Generative AI in the Workplace for 2025

To kick things off, Kevin Shepherdson, Founder & CEO of Straits Interactive, was joined for a lively discussion on the state of AI for 2025 and its implications for SMEs by Ang Yuit, President of ASME, Ms. Tin Pei Ling, Member of Parliament & Chairperson of Government Parliamentary Committee (GPC) for Digital Development and Information, Rajesh Sreenivasan, Partner and Head of the Technology, Media & Telecommunications Law Practice at Rajah & Tann Singapore LLP and Dr. Simon See, Global head of Nvidia AI Technology Centre.

From left to right: Kevin Shepherdson, Ang Yuit, Ms. Tin Pei Ling, Rajesh Sreenivasan, Dr. Simon See. 

From left to right: Kevin Shepherdson, Ang Yuit, Ms. Tin Pei Ling, Rajesh Sreenivasan, Dr. Simon See.

Shepherdson introduced the “6 Cs” that described the upcoming impacts of generative AI in the workplace for the year ahead. These are opportunities that SMEs should position themselves to take advantage of if they are aiming to leverage AI adoption for value-creation. At the same time, people in Governance, Risk and Compliance (GRC) roles would benefit from these learnings to inform AI management programs in their organisations. Here’s a condensed run-down of the 6 Cs for our readers.

1. Collection of Data

The price of more contextualised outputs is more proprietary data. As organisations start leveraging internal knowledge repositories to build AI models that reflect their unique expertise, accompanying measures in Data Governance and bias mitigation are to follow. Small Language Models (SLMs) are now a rising alternative to Retrieval-Augmented Generation (RAG) techniques for deriving insights from internal data using AI models, allowing companies to do this with greater computation efficiency and enhanced privacy.

2. Compute Power

Advances in NVIDIA’s Blackwell architecture, Amazon’s Ultracluster Supercomputer and Trainium Chips and cloud-based AI platforms are making cutting-edge AI accessible and scalable, particularly for SMEs. These innovations align with sustainability goals of environmental, social, and governance (ESG) nature, and democratise AI for smaller enterprises to tap on without prohibitive costs.

3. Context Window

The introduction of million-token long context windows in Google’s Gemini 1.5 models spells the start of AI's ability to remember multi-session conversational histories with an expanded - potentially infinite - memory. This will enhance workplace collaboration by maintaining project continuity between team members and delivering more contextually aware solutions. Advances in multimodal processing also means we can expect better synthesis and summary of complex documents to smooth out info-intensive communication.

4. Chain-of-Thought Reasoning

The emergence of strategic reasoning capabilities in AI through its ability to reason step-by-step is enabling it to tackle complex, multi-step challenges and decision-making with greater transparency and explainability in its outputs. First implemented in OpenAI’s o1 model and now its o3 successor, Chain-of-Thought prompt techniques are projected to be a critical workplace skill that improves trust in AI-generated outputs.

5. Customisation

From the advent of no-code platforms to autonomous AI agents, businesses are beginning to leverage tailored solutions that cater to specific industries, departments and workflows. We may see more employee-driven innovation of AI tools to automate menial tasks and deliver personalised support. The surfacing of the Chief AI Officer (CAIO) role also signals the growing need for leadership in AI strategy and governance to keep developments in line with organisational objectives.

6. Control

As AI becomes more autonomous and agentic in nature, there will be a demand for robust governance frameworks and real-time systems monitoring to mitigate risks and maintain ethical AI usage. The CAIO will have to drive interdepartmental collaboration to ensure accountable AI deployment that is in-sync with local and international regulations like the EU AI Act. This is important for maintaining user and stakeholder trust as we navigate new technologies.

Three Cs Join the Mix - Community, Cajoling and Collaboration

Expanding upon the foundational 6 Cs, our panelists chimed in with additional Cs that are just as crucial in successful AI adoption for 2025: Community, Cajoling, and Collaboration.

Community

Ms Tin championed the importance of community in cultivating a supportive ecosystem for SMEs. This includes attracting global talent, fostering cross-industry collaboration, and availing shared resources to empower SMEs. Talents come because they see opportunities in projects and real-life problems to solve. “And it's not just about the people coming together to try to cross-fertilise ideas, but also to pull resources [to support SMEs]”, she pointed out, citing existing initiatives like CTO-as-a-Service that help resource-constrained SMEs get off the ground with digital-readiness. 

Cajoling

For Sreenivasan, Singapore’s approach to governance is one of cajoling rather than controlling. It encourages businesses to experiment while staying ethical and compliant. “Singapore’s regulatory framework is about amplifying, clarifying, and enabling,” he explained. “Instead of legislating AI prematurely, we focus on engaging the community to understand what is going on.” He also advocates for collaboration with global counterparts and cross-talk between industries, stating, “Something like AI, which will not just impact a country in isolation but globally, requires collaboration. Through understanding the technology and building deep relationships with other parts of the world, we build bridges and systems that allow a degree of harmonisation.”

Collaboration

Ang called for a shift in the competitive mindset among SMEs to thrive in a market squeezed by global disruption. “Singapore, as a local market, is now a red ocean. We face intense competition not just from local players but also from international entrants bringing their own supply chains and software,” he commented. “It is more important, now than ever, to look beyond Singapore and collaborate.” Collaboration, Ang explained, is not just about working with peers but also about pooling resources across industries and nations. “How do we bring SMEs together to collectively solve challenges, open up new opportunities, and create value together?” he posed.

In the face of disruption, companies need to be adaptable too, says Ang. He advises SMEs to “be prepared to relook your processes and find new ways to serve your business and audience better.” To this end, Dr. See encouraged companies to capitalise on agentic AI to improve existing ways of working and client services. Especially when its combination with generative AI offers a more seamless path to crafting bespoke solutions.

But what would it take for employees with no technical expertise to develop capabilities in generative AI usage and customisation in the first place? Insights from the 2024 National AI Prompt Design Challenge paints a picture for us.

Our Prompt Design Champions Return to the Stage

Following the conclusion of the inaugural Youth and Open categories of our National AI Prompt Design Challenge last year, the winners from our Open category returned to the stage to share their experiences in a retrospective panel joined by key members of the judging team. Namely, Michael Low, Deputy Director of Curriculum Development & Digitalisation at SMU Academy, Harish Pillay, AI Verify Foundation, Ang and Shepherdson. The Challenge, which aligns with Singapore’s National Strategy 2.0 and the Digital For Life movement, sought to enable SMEs and students to ethically design their own use case-centric generative AI apps – with no coding knowledge. 

The competition was conceived by Straits Interactive and co-organised by ASME and Singapore Management University (SMU) Academy, along with leading bodies in the digital sphere, including AI Verify Foundation, Infocomm Media Development Authority (IMDA) and the Cyber Security Agency (CSA) of Singapore.

From left to right: Vernon Lim, Kevin Shepherdson, Ang Yuit, Michael Low, Harish Pillay. The discussion was moderated by Vernon Lim, AIFA Organising Chair. 

From left to right: Vernon Lim, Kevin Shepherdson, Ang Yuit, Michael Low, Harish Pillay. The discussion was moderated by Vernon Lim, AIFA Organising Chair.

Winning solutions from the six teams of the Open category addressed a range of organisational and social challenges. 

As business professionals who mostly identified as casual ChatGPT users, our champions stepped up to the task of learning prompt techniques, responsible app development, and deploying generative AI tools on our Capabara platform, all within a day’s time. The winner of the Most Innovative Award, Team AI Artisans, the only team with prior coding expertise, shared that to build and deploy their plant health diagnosis app from scratch would’ve taken at least a week. But the prompt training and no-code UI of the platform greatly expedited this process by making it layman-friendly. 

This lowered barrier to entry allowed the team from Focus on the Family Singapore to build their 1st place-winning app, the Virtual Parenting Companion. Armed with just English prompts and relevant industry experience. “The new programming language of tomorrow is human language,” states Shepherdson. 

From left to right: Daphne Tan and Gena Chua from Focus on the Family Singapore, who won 1st place for their Virtual Parenting Companion app.

From left to right: Daphne Tan and Gena Chua from Focus on the Family Singapore, who won 1st place for their Virtual Parenting Companion app.

Speed isn’t everything though. Pillay, the Chief Judge of the Challenge, stressed the critical role of testing in AI development, “You don’t just test once or twice—you test continuously.” According to him, some submissions failed because participants didn’t test for edge cases like adversarial prompts. On the contrary, Team NSIT’s IT Support Technical Advisor tool survived the judging panel’s best attempts to break into it, earning them the Most Responsible and Safe accolade.

The challenge also underlined the importance of ethical considerations. Some entries, such as a medicine recommendation tool and a tenancy agreement analyser, raised concerns due to the lack of domain expertise among their creators. “Just because AI allows you to do something doesn’t mean you should,” Shepherdson cautioned. Pillay reiterated this point, adding, “If you’re building an AI application, you must have a domain expert on your team.”

Beyond the competition, the Challenge was also about capability building. Low elaborated on the broader vision, where they’ve “designed a comprehensive roadmap for generative AI upskilling and reskilling.” Referencing the SMU-Straits Interactive learning roadmap for SMEs and individuals to adopt Generative AI, it offers pathways to either an Advanced Certificate in Generative AI, Ethics and Data Protection, or an Industry Practice Masters (IPM) Degree in Digital Economy - Data Protection and Data Governance. The opportunities are here for SMEs to upgrade themselves - they just need to take the first step.

Digital Transformation Needs Continuous Innovation

Two years on since the initial babble of fascination with a chatbot bearing conversational dexterity, the momentum of optimism lives on, carrying AI forth to its next evolution. For SMEs, this wave represents an opportunity—but only for those ready to embrace the challenge and step forward boldly.

Summing up the key message for SMEs, Ms. Tin urged businesses to adopt a forward-thinking mindset: "We cannot lose the mentality to keep innovating…If we keep thinking that we are just here to learn how to use [AI tools], we end up being users of the software rather than innovators and value builders." She nudged to take bold steps into the AI era, adding, “While there are risks, let us also be bold and hopeful. Just go out there, try it out, and you never know where this will lead.”

Capabara, our Next-Gen AI Capability-as-a-Service platform, is currently available on a limited trial basis. If your company is interested in developing generative AI competencies and capabilities with us, write in to sales@straitsinteractive.com with your company email address. You can also stay tuned to Capabara’s latest developments by following CAPABARA on LinkedIn or heading over to capabara.com to find out more about how your organisation can be empowered through safe, secure and sustainable generative AI adoption.

This article was first published on The Governance Age on 28 Jan 2025.