TAKEAWAYS
“AI-enabled”. The phrase the National University Health System (NUHS) uses to describe its AI ambition is deliberate. Not “AI-powered”, not “AI-first”, but “AI-enabled”. The distinction matters.
“AI augments – it doesn’t replace,” says Wong Soo Min, CA (Singapore), Group Chief Financial Officer of NUHS. “The core of healthcare still lies in human interaction.”
In healthcare, where regulatory requirements, patient safety, and ethical accountability are paramount, human oversight remains non-negotiable. AI, therefore, is not an end in itself, but a tool guided by organisational values. Speaking at a professional sharing session organised by the Chief Financial Officer Committee (CFOC) of ISCA, Ms Wong offered an inside-out view of NUHS’ measured and principled AI journey.
NUHS’ AI journey has progressed through clear, structured phases.
Supporting this roadmap are five foundational pillars: leadership and governance, technology infrastructure, capability development, culture, and investment discipline.
As Ms Wong emphasises, infrastructure is non-negotiable. NUHS has spent several years building a scalable, integrated data platform and cleansing data drawn from multiple disparate source systems. The work is laborious and often invisible, but essential.
“Without the right infrastructure and system integration,” she notes, “there is no way to tap on data or run AI reliably.”
Recognising that AI literacy cannot be taken for granted, NUHS has implemented a tiered AI competency framework covering all staff levels from frontline employees to senior leadership. The framework spans four levels, from basic awareness to advanced AI engineering.
This investment in structured upskilling reflects a clear belief, that even the most sophisticated AI tools will underperform if users lack confidence, understanding, or trust in how they work.
For an organisation in growth mode – with one new hospital being built and two others undergoing redevelopment – AI is not about reducing headcount but about enabling staff to operate more effectively amid rising care demand and workforce constraints.
On investment discipline, Ms Wong was notably candid about the realities of prioritisation and trade-offs.
AI benefits, particularly in healthcare, often materialise only after an extended period. Solutions that affect clinical quality or patient safety require rigorous validation and governance. As such, NUHS places strong emphasis on prioritising proof-of-value initiatives, scaling only where impact is demonstrated, and are prepared to walk away when value fails to justify continued investment.
Equally important are decisions around what to build internally versus what to procure commercially.
The structural challenge remains familiar to CFOs everywhere: understanding the potential returns of AI investments long before payback is visible. Outcomes such as lower hospitalisation rates, improved chronic disease management, and reduced emergency demand are meaningful, but often difficult to attribute to a specific technology investment made years earlier.
The discipline required, Ms Wong suggests, is not merely financial. It is the conviction to commit to long-horizon outcomes aligned with mission and public good.
For finance professionals, NUHS’ experience offers practical and transferable lessons.
During the pandemic, the finance team developed data analytics tools using Python and Visual Basic for Applications (VBA) to automate compliance reviews. These tools consolidated fragmented reports from outsourced service providers and automatically identified instances of non-compliance that informed payment decisions – eliminating hours of manual work previously required.
This groundwork paved the way for the deployment of an in-house GenAI chatbot to handle policy queries and case-based costing questions. Previously, such queries landed directly in finance inboxes, consuming time better spent on higher-value analysis. Adoption, however, required more than technical deployment; leadership had to actively reinforce self-service as the norm.
NUHS has also built and deployed a machine-learning model to predict potential patient payment defaults. The objective is not enforcement, but early support – flagging cases where social workers or financial counsellors can intervene proactively.
This year’s major finance initiative is exploring an AI-enabled financial close, including the use of GenAI to generate variance analysis and management narratives.
What emerges from Ms Wong’s account is not a story of AI deployed for signalling or novelty, but one rooted in realism, governance, and purpose. NUHS is using AI to better serve both staff and patients amid structural challenges – an ageing population, rising chronic disease burden, and a constrained healthcare workforce.
For healthcare finance professionals, the message is clear: the finance function cannot sit at the periphery of AI transformation. Data quality, investment prioritisation, cost modelling, and the quantification of clinical outcomes are ultimately finance responsibilities.
The CFO’s role in an AI-enabled health system is therefore not simply to fund transformation, but to lead it – with discipline, clarity, and a steadfast commitment to keeping people at the centre.
