NCS says the main barriers preventing AI experiments from reaching large-scale production aren’t confined to technology. Costs, governance, unchanged business processes, skills gaps, and employees’ fear of displacement must also be addressed. To help enterprises make that transition faster, NCS is itself becoming an organisation in which AI plays a central role in delivering technology services.
During the NCS AI Impact 2026 tech forum, CEO Sam Liew and Chief AI Officer Edward Chen explained the components of the company’s enterprise AI strategy, including practical applications in healthcare, education, and transport.
Three-part mission
Liew began his keynote by outlining three priorities for NCS: reorganising the company as an AI-led technology firm, releasing the NCS AI Playbook, and equipping its workforce with AI skills.
Liew said customers no longer need only someone who can implement technology. They need a partner that can coordinate entire ecosystems and apply rapidly developing AI technology to practical needs.
“(That’s why we are) building a stronger industry focus, so that our teams understand not just technology, but also the realities of our clients’ sectors. It means a new operating model in which AI helps teams move from idea to implementation in days and weeks, not months and years,” he said.
Liew said every NCS employee, including those in corporate services, is now paired with at least three role-based AI agents. One business support manager, who has worked at the company for 27 years but is neither a developer nor technically trained, now uses NCSgpt for sales registration and project operations.
“They configured custom prompts in NCSgpt and built an AI assistant trained on NCS policy guidelines, turning tedious, manual project scope reviews into an automated workflow. The team can therefore use standardised workflows that comply with company requirements, with AI support available 24/7,” Liew said.
Meanwhile, the NCS AI Playbook is a blueprint based on practitioners’ experience that provides organisations with strategies, frameworks, and execution processes for turning AI investments into measurable business value at scale. Liew said it draws on lessons from implementations in Singapore, Hong Kong, Australia, and China. Its main conclusion is: “We need technology and people, and everything in between, to be ready.”
Liew also listed several reasons for his assertion that 95% of AI pilots never reach production: unoptimised AI costs, unclear governance, unchanged business processes, delays in reskilling employees, and employees’ fear of being displaced.
“The core message is simple: AI has to work in each client’s actual operating environment. Our AI Playbook is a practical guide to help organisations turn their AI ambitions into AI outcomes,” he remarked.
Liew also said NCS employees were his top priority and committed to continuing to invest in the workforce and recruit new talent.
“We have three work-study programs that allow employees to upgrade their qualifications while working full-time with NCS: from an ITE qualification to a diploma, from a diploma to a degree, and from a degree to a Master of Technology. Across all the programs, we have onboarded more than 1,900 work-study employees,” he said.
Scaling AI
Examining why organisations’ AI projects fail, NCS Chief AI Officer Edward Chen said most people are taking the right steps, but only at a superficial level.
“Chatbots, summarisation tools, and automated workflows deliver real productivity gains. However, when AI is layered on top of the way you already work — with the same process, approval chain, and operating model — you only make the old engine faster. You do not build a new engine,” he pointed out.
To address this problem, Chen announced five additions to NCS’s Sunshine.AI suite:
- Sunshine.core, the underlying platform for building and operating AI agents in production.
- Sunshine.builder, an application that lets business analysts use AI to build software without writing code.
- Sunshine.chilliclaw, an AI assistant that adds agent-based capabilities to software employees already use.
- Sunshine.commanderAI, a physical AI platform that connects AI with robotics and provides a single command centre for robot fleets from multiple vendors.
- Sunshine.guardian, a safety and assurance system for agent-based AI in production.
“Across every enterprise we work with, demand for AI is coming from the people you would never expect: operations leads, project managers, and business analysts. They have real problems and real ideas, but never enough engineers to build what they need,” Chen said.
Practical applications
NCS has worked with organisations in several industries to expand the use of AI within enterprises. In healthcare, NCS signed a memorandum of understanding with IHH Healthcare to establish a joint AI centre of excellence. The centre will jointly develop and deploy AI systems across IHH’s multinational network to support clinical practice, operational efficiency, and cost management.
NCS is also deploying agent-based AI in pilot areas within NHG Health, including biometric identification, digital pathology, and human resources transformation, to simplify operations, support clinical workflows, and improve patient outcomes.
In education, NCS is deploying an AI tutor at Ngee Ann Polytechnic to address learning gaps. NCS said adult learners in a recent pilot reported that the tutor had a positive effect. The company is also incorporating its Sunshine.coder tool into the institution’s information and communications technology curriculum to train people to work with AI.
In transport, NCS is building an autonomous shuttle service for employees of South Korean company Autonomous A2Z. The project integrates the ROii vehicle with NCS’s RobotManager platform to develop an operating model for future client projects.
“The real opportunity isn’t incremental improvement; it’s redesigning core operations to achieve much larger results,” Liew said. “To deploy AI at scale, organisations need partners that combine industry knowledge with sovereign platforms designed for enterprise use and built to meet practical governance and security requirements. They also need a strong ecosystem behind them.”
