SINGAPORE – Plans are afoot to push the use of
artificial intelligence
across four key sectors on a national scale to unlock Singapore’s competitive advantage.
Prime Minister Lawrence Wong said in his Feb 12 Budget 2026 speech that he will front a new inter-ministerial committee to drive the execution of
AI missions
in advanced manufacturing, finance, healthcare, and connectivity and logistics.
To fully realise AI’s potential, Singapore needs to move beyond individual pilots and isolated experiments, said PM Wong. “We must organise at a national level, and move with speed and scale,” he said.
The Straits Times examines the current AI projects in these four sectors, where the
gaps in deployment
are and what a national level of implementation in each of these sectors might look like.
AI is often used for predictive maintenance to forecast when machines are likely to fail so repairs can be carried out before breakdowns disrupt production.
Predictive maintenance systems typically analyse real-time data from sensors, tracking factors such as vibration, temperature and equipment performance to identify patterns that indicate potential faults.
Professor Stefan Winkler from Singapore Institute of Technology’s (SIT) engineering cluster said: “These tools are sometimes deployed for individual machines, or perhaps a production line, rather than across entire supply chains, limiting productivity gains.”
For example, semiconductor foundry GlobalFoundries deploys the technology in its 300mm semiconductor fabrication plant in Woodlands.
Separately, A*Star and the Ministry of Trade and Industry have been operating a Sectoral AI Centre of Excellence for Manufacturing since September 2024 to help the industry tap AI for purposes such as predictive maintenance, product design and industrial automation. Thirteen companies, including drinks producer Coca-Cola and tech manufacturer Philips, were working with the centre at launch.
Mr Zavier Wong, market analyst at Israel-headquartered online trading platform eToro, said that machines that perform predictive maintenance can be linked to systems that manage production scheduling and the procurement of spare parts.
So even before a manufacturing plant breaks down or receives scheduled maintenance, other plants can automatically take over the production. A linked procurement system can also automatically order the necessary spare parts for repair.
Currently, many of these systems exist in isolation and human reviewers are manually ordering parts or shifting production work to other facilities.
AI applications for route optimisation or demand forecasting are increasingly common, helping companies move goods more efficiently.
Route optimisation finds the fastest or least costly delivery paths, while AI-powered demand forecasting predicts future customer needs by analysing past sales and trends to plan inventory and shipments to match demand without over- or under-stocking. Logistics giants DHL Express Singapore and FedEx Singapore use such technologies.
Mr Vivek Lath, a partner at McKinsey & Company, said: “For instance, AI may predict the estimated time of arrival shipments, but if this data and model isn’t integrated with trucking schedules or warehouse staffing or information to end consumers.”
For a start, live traffic data on road closures, accidents or heavy traffic can be fed into fleet management systems to divert vehicles to minimise delivery delays. The fleet management system can also be linked to customer service platforms to automatically update customers about potential delays.
Freight forwarders, warehouses, ports and delivery firms can also be connected through real-time data sharing to allow certain decisions to be made by AI systems without human intervention. These decisions include how goods can be packed in a cargo container to maximise storage, and the verification of shipping documents for automated customs clearance.
Mr Manik Bhandari, co-chair of trade association SGTech’s AI, cloud and data chapter, said that challanges include cross-border compliance. Common data standards also need to be established for data sharing.
Experts said that the financial sector is the most mature in terms of AI adoption.
Mr Varun Arora, managing partner for South-east Asia for United States-headquartered consulting firm Kearney, said that the sector has large, well-organised transaction data, highly digitalised operations, and strong governance frameworks in regulated environments.
AI is currently deployed to provide customer profiling and service, fraud detection and risk management, and in automating back-office tasks like transaction processing, reporting and regulatory compliance.
At UOB, for instance, its banking app UOB TMRW uses AI to send deal recommendations based on customers’ spending and saving habits. The bank also uses AI to detect fraud and money laundering.
In its 2024 annual report, OCBC said that six million decisions in the bank are made by AI on a daily basis, from personalising credit card deals for customers to fraud detection. An AI model is also used to identify customers who require assistance with their home loans, allowing the bank to engage them proactively, resulting in a 20 per cent increase in customer retention.
McKinsey’s Mr Lath said that AI can be integrated from onboarding, risk assessment to compliance.
For example, AI models can identify if a customer can be pre-approved for a home loan after drawing insights from his income, spending and credit history. If the customer clicks on the pre-approved offer, his application form can automatically be filled. If the loan is taken, the AI model can continue to monitor repayments and debt levels to detect early signs of financial stress and intervene promptly with offers to restructure the loan.
But there are challenges to ensure that such AI decisions are transparent and accountable, eToro’s Mr Wong said. For instance, firms need to explain how a recommendation or a loan decision was made.
Mr Nigel Lee, country general manager for Lenovo Singapore, said that customers must also be informed when algorithms influence financial decisions.
Singapore General Hospital and Changi General Hospital have deployed an AI system that can spot heart and lung abnormalities from chest X-rays.
Singapore national health technology agency Synapxe’s chatbot, called Lab Report Buddy, can analyse medical lab reports and translate medical jargon for patients.
Clinicians across some hospitals also readily rely on AI transcription and diagnostic tools to summarise medical records and analyse medical images.
SIT’s Prof Winkler said that strict data-privacy requirements in the healthcare sector and the complexity of hospital workflows have resulted in many standalone AI implementations.
He said opportunities lie in having patient record updates automatically triggering follow-ups or specialist referrals, urgent tests scheduling, hospital discharge coordination and long-term patient monitoring. AI tools can also help with staff rostering, ensuring clinicians are available when they are needed.
But AI diagnosis needs to show high accuracy rates consistently before more automation can occur.
eToro’s Mr Wong said: “There’s very little margin for error… unlike other sectors where mistakes may lead to financial loss or operational disruption, the stakes here are fundamentally higher because they carry direct human consequences.”
The
passing of the Health Information Act in Singapore on
Jan 12
has laid the foundation for more integrated AI deployments across the healthcare sector.
The new legislation, which is slated to come into effect by early 2027, requires all public and private healthcare providers to contribute patients’ health information to a national repository called the National Electronic Health Record (NEHR) system. By sharing patients’ data, it is hoped that patients need not have to repeat medical tests or bring along physical scan reports when they change providers.
Health Minister Ong Ye Kung had previously said that anonymised data from the NEHR can be used to train AI models to deliver predictive, preventive care.