Singapore enterprises are investing billions into AI, with the latest investments going into agentic AI: Systems that autonomously orchestrate, optimise, and execute complex business processes without the need for constant human intervention.
From managing supply chains and regulatory compliance to automating customer service and internal workflows, these agents promise to reduce the burden on human decision-making.
Yet, according to Gartner, nearly 40% of these projects are expected to stall or be cancelled by 2027, a sobering statistic for technology currently riding high on boardroom hype. The root cause continues to be fragmented data. This same old villain is still the main reason ambitious AI initiatives grind to a halt before they ever reach production.
To bridge the gap between AI’s potential and business outcomes, Singapore’s next wave of digital transformation needs a stronger foundation in data-centric infrastructure. Data models that can connect and contextualise information provide the context, memory, and traceability required for AI to deliver sustained returns, whether through cost savings, improved efficiency, or new growth opportunities.
Pitfalls and promises
Backed by strong government initiatives such as the SG$150 million Enterprise Compute Initiative and SMEs Go Digital program, Singapore is among Asia-Pacific’s (APAC) leading AI adopters. Analysts forecast that the APAC agentic AI market will reach US$110 billion by 2028, with NCS committing SG$130 million over three years to regional AI transformation.
But if agentic AI is so promising, why aren’t the profits matching the hype?
The answer is simple, but the execution isn’t. Implementing agentic AI at scale is tougher than many first expect. Gartner’s 2025 Hype Cycle captures it perfectly, where agentic AI sits at the “Peak of Inflated Expectations,” a classic scenario where enthusiasm races ahead of reality. It’s not that agentic AI cannot deliver, but too many organisations underestimate the operational complexities. From legacy systems to messy data, and a still-maturing talent pool, many organisations are discovering that the road from proof of concept to return on investment (ROI) is far bumpier and longer than expected.
So why keep investing despite these frustrations?
It’s the promise that once agentic AI moves past its growing pains, it will fundamentally reshape what businesses can achieve. However, turning a promise into profits takes more than buzzwords; it requires the right infrastructure.
Building the right foundations for ROI
Data models provide the memory, context, and relationships that enable AI systems to act with greater intelligence. These approaches don’t just store data; they enable relationships to be drawn between previously isolated systems, processes, and people.
Think of it as giving AI a way to understand business context: the connections between customers, products, employees, and operations. With the right foundations, AI agents can access past experiences, make sense of the present, and act with greater intelligence.
In Singapore, businesses are beginning to invest in these capabilities to surface insights, enhance traceability, and drive smarter decisions across sectors from banking and telecoms to public services such as talent and career management.
Developing this kind of second layer in their infrastructure allows businesses to move beyond agentic AI’s initial hype and ground it towards productive outcomes.
Moving from short-term expectations to long-term foundations
So, where do we go from here? For CIOs and technology leaders, the question is no longer whether to invest in AI, but how to invest in it. Should the next dollar go toward more LLM API calls, another chatbot subscription, or toward wiring the business with the data infrastructure it actually needs to deliver results?
The answer lies in shifting focus, from short-term pilots to long-term foundations. It’s not about chasing the next shiny AI trend, but about putting the groundwork in place to support meaningful, scalable progress. Singapore has the policy support, the tech-savvy workforce, and the ambition to lead this next chapter; so long as we invest in stronger data foundations, and not just bots.
