Many small business owners in Singapore are keen to adopt artificial intelligence (AI), but putting it into practice remains a challenge. A survey commissioned by Meta and conducted by YouGov found that 94% of Singapore small business owners surveyed wanted more guidance on using AI, while 37% wanted clearer information on costs and returns on investment. Another 57% saw cost as a potential major barrier to businesses taking part in AI training.
Meta Singapore managing director Nicole Tan, who took up the role in June last year, shared the findings at Meta Upskill Day on Monday (16 November).
The survey also found that 82% of small business owners surveyed who were already using AI had seen productivity improve over the past 12 months. Of these, 62% used the time saved to improve the quality of their existing work.
In an interview with Lianhe Zaobao, Tan said small and medium-sized enterprises (SMEs) could start adopting AI by establishing online channels, then continually testing and learning to identify metrics suited to their businesses.
With limited resources and funds, how can small businesses take the first step towards using AI? Huang Weijian, co-founder of Singapore communication training firm SpeakEasy Lab, told Lianhe Zaobao that while concerns about disappointing returns on investment were natural, businesses would not know which methods worked without first trying them.
He stressed that experimenting did not mean investing blindly. Businesses should draw up a plan, such as testing different marketing content and formats over a set period and comparing audience responses to find suitable ways to promote themselves.
The company estimates that about 95% of enquiries from new customers come through Meta’s social media platforms and advertisements. “For a small business like ours, every enquiry matters.” AI customer service has reduced response times for routine enquiries from about an hour to almost instant replies, saving 10 to 15 hours a week.
At Monday’s roundtable discussion, Low Yen Ling, Singapore’s Senior Minister of State for Culture, Community and Youth, and for Energy, Trade and Industry, encouraged SMEs to embrace the “3As” of AI adoption: assist, automate workflows, and augment products and services.
She also stressed that business transformation must go hand in hand with workforce transformation. Even a neighbourhood shop with just two or three employees could, with AI and agentic technology, gain the capabilities of a team of 10 or more.
“Who says neighbourhood shops can only do business locally?”
Low also said neighbourhood shops need not confine themselves to physical premises or the Singapore market. Many local brands that have expanded internationally also began as neighbourhood shops.
She said neighbourhood shops could serve as incubators for testing products and use AI and digital platforms to grow their businesses and enter overseas markets. With Singapore taking over the ASEAN chairmanship next year, neighbourhood businesses should seize opportunities in the region.
Businesses must take responsibility for AI-generated advertising
As AI-generated advertising becomes increasingly common, Meta has launched Dream Render, an AI video creation tool that lets users generate promotional videos by entering text descriptions and uploading product and brand photographs.
However, some consumers are concerned about misleading advertisements, making them less willing to buy and potentially damaging brands’ reputations.
Responding to these concerns, Tan stressed that businesses must ensure they have the rights to use the relevant materials and take responsibility for ensuring their content complies with the rules. Consumers can also report misleading content to the platform.
Push Sharma, founder of mint brand Haldy, had also questioned the originality and authenticity of AI content. But he believes the key to building trust in AI advertising is whether the content genuinely reflects the brand.
He said the company has people check, edit and approve AI videos. It also shows them to a segment of its audience for feedback before publishing them, and labels the content as AI-generated.