AI Guide For Boards in Singapore

AI Guide for Boards in Singapore

In collaboration with

AI Guide for Boards in Singapore

Contents

Foreword

Preface

Our Partners

Executive Summary

SECTION I: The Board Mandate Chapter 1: Why Boards Must Lead 1.1 A Primer for Directors 1.2 Key Shis in the AI Industry 1.3 Value Creation and Disruption 1.4 Resilience in a Disrupted Economy Chapter 2: The Case For (and Against) AI 2.1 Dening Strategic Ambition

14 16 18 23

24 25 26 27 28 30 32 33 34 35 36

2.2 Industry-Specic Perspective 2.3 Cross-Function Development

2.4 Emerging Disruptors 2.5 Potential Challenges

Chapter 3: The AI-Ready Board 3.1 Board Roles and Responsibilities 3.2 Assessing Board Capabilities 3.3 Enhancing Board Composition

3.4 Committee Structures 3.5 Continuous Learning

3.6 Building an AI-Ready Board

SECTION II: Strategy and Value Creation Chapter 4: Integrating AI into Corporate Strategy 4.1 Identifying AI Value Pools

38 39 40 41

4.2 Assessing Competitive Advantage

4.3 Resource Allocation

4.4 Change Management and Monitoring

Chapter 5: Measuring AI Value 5.1 Establishing AI-Specic Performance Metrics

43 46 47 48

5.2 Measuring Return on Investment 5.3 Dening Success Criteria 5.4 Embedding Accountability

Section III: Governance and Oversight Chapter 6: The Regulatory Landscape 6.1 Singapore’s AI Governance Approach

52 55

6.2 Cross-Border Dynamics

Chapter 7: Board-Led AI Governance 7.1 AI Governance Framework 7.2 Integrating AI into Corporate Governance 7.3 Integrating AI into Enterprise Risk Management

57 62 63 65 67 69 70

7.4 Dening Roles and Responsibilities 7.5 Establishing AI Governance Policies

7.6 Governing the AI Lifecycle

7.7 Regulatory Oversight and Reporting

Section IV: Ethics and Trust Chapter 8: Responsible AI and Building Trust 8.1 Responsible and Ethical AI

74 76 77 80

8.2 Fostering Responsible AI Culture 8.3 Ecosystem Risks and Governance

8.4 Stakeholder Governance

Section V: Organisational Readiness and Capability Chapter 9: Ensuring AI and Digital Resilience 9.1 Digital Resilience at the Core 84 9.2 Scaling AI 85 9.3 Assessing Digital Readiness 86 9.4 Ensuring Data Integrity 87 9.5 Embedding AI 88

Chapter 10: Operationalising Digital Resilience 10.1 Infrastructure Robustness

90 91 93 94

10.2 Cyber Security Resilience 10.3 Operational Resilience 10.4 Model Risk Management

Chapter 11: Talent and Technology 11.1 Building AI Talent

95 96 96 99

11.2 Data Strategy

11.3 Security and Accessibility 11.4 Model Lifecycle Governance 11.5 Monitoring and Assurance

101 102

11.6 Intellectual Property

Section VI: Operationalisation Chapter 12: Strategic Roadmap for Boards 12.1 Operationalising AI

104 108

12.2 Board Stewardship

Acknowledgement

109

Appendices 1.A e AI Lexicon 1.B e Dual Lens of Governance: Conformance and Performance 1.C AI-Enabled Productivity 1.D Scaling Expertise and Customer Value 1.E Stewardship of People Resources 2.A Incremental Productivity vs Transformational Change 3.A Board Oversight of AI

3.B Board vs Management Responsibilities 3.C Board AI Competency Framework 3.D Strengthening Board Capability for AI 3.E Key Functions of an AI-Ready Board 4.A A Value Pool Approach

4.B Core Pillars of AI Investment 4.C e AI Integration Iceberg

5.A Stages of Maturity in the AI Lifecycle 5.B AI Performance Metrics: Examples 6.A Singapore as a Trusted AI Hub 6.B Singapore’s Approach to AI Governance 6.C Monetary Authority of Singapore: FEAT Principles 6.D AI Verify Testing Framework 6.E AI-Related Regulations, Frameworks and Tools: Examples 7.D Key Risk Indicators: Examples 7.E Specialised Board Committees 7.F Tiered Risk Hierarchy: Example 7.G Reporting on AI Performance and Value 7.H Board Checklist for AI Governance 8.A Stakeholder Trust Expectations 8.B Four Core Principles in Building a Responsible AI Culture 8.C AI-Specic Due Diligence 8.D Emerging AI Risks 8.E Board Checklist for Stakeholders 7.A A Risk-Tiered Approach 7.B Illustrative Risk Scenarios 7.C Board Oversight Process

9.A Digital Readiness Checklist 10.A Oversight of AI Infrastructure 10.B Secure AI Deployment 11.A e Build-Buy-Partner Framework 11.B Data Sensitivity Classication: Examples

11.C Model Lifecycle Governance 11.D Monitoring for Assurance 11.E AI-Related IP Risk Landscape 12.A Seven-Step Strategic Roadmap

AI GUIDE FOR BOARDS IN SINGAPORE

Foreword

Stewarding continuous transformation

Every generation of boards is dened by the strategic shis it must steward.

is, I believe, calls for three important shis in board stewardship.

Today, articial intelligence (AI) is one such shi – but it is also something more.

e rst shi is from governing AI to governing continuous transformation.

It is changing not only how organisations operate, but how they compete, create value and adapt to change. Unlike previous technology waves, AI is evolving at unprecedented speed. New models, new capabilities and entirely new ways of working are emerging not over years, but months, even weeks. Technological change is accelerating. As AI becomes more accessible, competitive advantage will increasingly come not from technology alone but from an organisation’s ability to learn, adapt and continually create new value. e true measure of transformation is not the value it creates today, but whether it strengthens the organisation’s ability to create value tomorrow.

Boards are no longer overseeing AI initiatives. eir responsibility is to ensure the organisation continually strengthens its ability to learn, adapt and transform – as AI itself continues to evolve. Sustainable value comes not from deploying more AI tools, but from redesigning work, operating models and ways of working so that people and AI create value together.

e second shi is from managing risk to building risk intelligence.

Every technological breakthrough expands both opportunity and risk. AI is no exception. It introduces new challenges, from cyber security and data governance to model integrity, bias and accountability. Yet, it also provides powerful new ways to anticipate, understand and manage those risks. Boards should foster a culture of risk intelligence, where management can identify emerging risks early, distinguish risks worth taking from those that must be mitigated, and leverage AI itself to strengthen resilience, governance and trust. As we recognised in the SID Cyber Resilience Guide for Boards in Singapore , resilience must be designed into our digital systems. In the

I believe this is now the dening responsibility of board stewardship.

Instead of asking: “ How can AI help us work better? ”, boards should ask “ How can AI make our organisation better? ”

at is fundamentally a question of governance.

e governance challenge is larger than AI itself. It requires boards to rethink how they steward organisations through continuous change.

AI GUIDE FOR BOARDS IN SINGAPORE

AI era, adaptability must also be designed into our organisations.

of perspectives – from technology and cyber security to digital transformation, organisational change and human capital – to ask better questions, challenge assumptions and guide organisations through continuous transformation. is AI Guide for Boards in Singapore is timely. It provides directors with practical frameworks to navigate these shis with condence.

e third shi is from stewarding capital to stewarding organisational capability.

Governance for an AI era

AI is not simply another technology investment.

Technology creates possibilities.

Governance in an AI era is no longer about keeping pace with technology.

Organisational capability is what turns those possibilities into enduring advantage. Every investment in leadership, people, governance, data, innovation and continuous learning should therefore do more than improve performance today. It should strengthen the organisation’s ability to seize tomorrow’s opportunities. Boards should ask themselves: • Has this transformation made us stronger? • Will it make the next transformation easier? • Are we better able to create value tomorrow? Organisations that strengthen these capabilities become better at recognising opportunities, responding to change and creating value over time. at is an advantage competitors will struggle to replicate.

It is about ensuring every transformation leaves the organisation stronger than before.

e organisations that thrive will not simply be those that adopt AI successfully.

ey will be those that become stronger because they continuously learn, adapt and evolve. e board's role is therefore no longer simply to oversee technology. It is to ensure every technological change strengthens the organisation's long-term ability to compete and create value. at, I believe, is now the dening stewardship responsibility before every board.

Tan Kiat How Senior Minister of State Ministry of Digital Development and Information Ministry of Health

is also requires boards themselves to evolve.

Board renewal is not simply about adding technology expertise. It is about ensuring the board collectively brings a diversity

AI GUIDE FOR BOARDS IN SINGAPORE

Preface

Articial intelligence (AI) has quickly moved from emerging technology to become a strategic priority for every boardroom. Across industries, organisations are exploring how AI can improve productivity, enhance decision-making, create new business opportunities and strengthen competitiveness. AI raises important questions about governance, accountability, ethics, cyber resilience and organisational readiness. Boards are expected to provide oversight and strategic direction on AI, yet the pace of technological advancement can make it dicult to distinguish between short-term hype and long-term value. Directors do not need to become AI experts. ey do, however, need to understand the implications of AI, ask the right questions, guide management eectively and ensure that AI is deployed responsibly and sustainably. is is why the Singapore Institute of Directors has developed the AI Guide for Boards in Singapore . Building on Singapore’s ambitions under the National AI Strategy 2.0, this Guide is designed to help directors understand their role in overseeing AI adoption and transformation. It provides practical frameworks, governance considerations and boardroom questions that can support more informed discussions and decision- making.

We are grateful to our knowledge partners, contributors and reviewers who have generously shared their expertise and experience. eir collective insights have helped ensure that this Guide remains practical, relevant and grounded in the realities faced by boards today. Responsible governance and innovation are not competing objectives. Organisations that establish clear governance frameworks, build trust and strengthen organisational capabilities will be better positioned to capture the benets of AI at scale. Eective governance is an enabler of sustainable innovation. As directors, we have navigated previous waves of technological disruption, from digitisation and cloud computing to cyber security and digital transformation. AI represents the next chapter in that journey.

While the opportunities are signicant, so too are the responsibilities.

We hope this Guide serves as a practical companion for directors as they navigate these opportunities and challenges. More importantly, we hope it encourages meaningful boardroom conversations about how AI can be harnessed responsibly to create enduring value for organisations, stakeholders and society.

Yeoh Oon Jin Chairman Singapore Institute of Directors

AI GUIDE FOR BOARDS IN SINGAPORE

Our Partners

Infocomm Media Development Authority

As the AI landscape continues to evolve and new risks emerge, trust in organisations is paramount. Boards have a critical role to play in ensuring that corporates are AI ready, balancing innovation enablement with responsible oversight. is AI Guide for Boards in Singapore therefore serves as a useful reference for directors looking to better understand how they can carry out their responsibilities eectively to realise AI’s transformative potential. To that end, IMDA supports the Singapore Institute of Directors’ eorts to strengthen boards’ AI capabilities as we build an AI-bilingual workforce to power Singapore’s digital economy ambitions.

Ng Cher Pong Chief Executive Ocer, Infocomm Media Development Authority

Microsoft

Microso is delighted to collaborate with the Singapore Institute of Directors on this guide because AI is fast becoming the next “general purpose technology” that will profoundly reshape industries, businesses and jobs. AI transformation is hence now a boardroom responsibility and a strategic capability. As a longstanding contributor to Singapore’s digital transformation, Microso brings global responsible-AI expertise, enterprise transformation experience, and a deep commitment to Singapore’s digital economy. By working alongside SID, IMDA and ecosystem partners, we aim to provide practical guidance that helps boards strengthen trust, accelerate innovation, and turn AI ambition into measurable outcomes for local Singapore organisations.

Chia Wee Luen Managing Director for Singapore, Microso

OpenAI

AI is not a technology agenda, it is a leadership test. It is already reshaping how companies compete, make decisions and create value, and the greatest risk for many organisations may no longer be moving too fast, but waiting too long. Boards must do more than oversee AI: they must guide the ambition, challenge management to turn experimentation into enterprise-wide impact, and ensure innovation advances with trust and accountability. Singapore has an opportunity to lead this next era – not simply by adopting AI, but by dening what responsible, high-impact deployment looks like. At OpenAI, we are proud to contribute to this guide and to work alongside the leaders shaping that future.

Andy Brown Head of Go-To-Market APAC, OpenAI

AI GUIDE FOR BOARDS IN SINGAPORE Executive Summary

Executive Summary

e AI Guide for Boards in Singapore is tailored specically for board directors in Singapore. Published by the Singapore Institute of Directors (SID), the guide provides an operational and strategic blueprint to lead organisations and management teams in successfully leveraging articial intelligence (AI).

Key highlights

e guide is systematically divided into ve foundational pillars, culminating in a Seven- Step Strategic Roadmap . • e board mandate (Section I): Sets the tone at the top, charting the evolution of AI into generative and agentic systems. It highlights the shi from hype to disciplined return on investment, focused on measurable business outcomes through practical implementation. • Strategy and value creation (Section II): Outlines how boards must challenge management to pursue both “best- in-class” (incremental productivity) and “only-in-class” (transformational business models) strategies while securing proprietary data moats. • Governance and oversight (Section III): Proposes potential board-level oversight through clear policies (dening intent, rules and processes) and comprehensive risk classication inventories. • Ethics and trust (Section IV): Focuses on mitigating algorithmic bias, hallucinations and expanding cyber-attack surfaces. It advocates a proportionate governance approach tiered by AI impact (productivity, internal processes, or customer-facing).

Key objectives

e primary objective of this guide is to elevate the duciary literacy of board members, shiing AI oversight from a passive tech-committee topic to a core boardroom mandate. It aims to bridge the AI knowledge asymmetry between boards and management, transforming directors into active stewards capable of balancing conformance (risk management) with performance (strategic growth). is guide aims to provide practical and localised frameworks and considerations. It directly aligns corporate governance with Singapore’s National AI Strategy 2.0 (NAIS 2.0), positioning local enterprises to leverage national initiatives like the IMDA Model AI Governance Frameworks. By oering actionable frameworks, the guide helps organisations escape “pilot purgatory” and build scalable AI platforms that protect (and grow) enterprise value and cultivate stakeholder trust.

AI GUIDE FOR BOARDS IN SINGAPORE

Executive Summary

• Organisational readiness and capability (Section V): Explores building an AI-ready board via upskilling and skills matrices. It emphasises stringent due diligence for third- party AI, noting that outsourcing technology does not outsource responsibility.

e Guide positions robust AI governance not as a brake on innovation, but as a resource to support boards in leading and creating an operating model that makes AI adoption scalable and sustainable for their organisations.

Seven-Step Strategic Roadmap

Section VI provides directors with a strategic roadmap to supervise the operationalisation of AI:

1. Evaluate strategic alignment: Classify projects across a three-horizon view and solve the capital-ecient “buy vs build” dilemma. 2. Develop governance structures: Design and implement a suitable governance strategy, adopt Singapore’s Model AI Governance Frameworks and establish clear boundaries for human oversight. 3. Strengthen data infrastructure: Ensure legal “right to use” data in compliance with regulatory requirements, while monitoring high-compute carbon footprints for SGX climate disclosures. 4. Establish risk management guardrails: Mandate adversarial testing and identify “hard red lines” where automated decisions are prohibited. 5. Monitor adoption and performance: Review AI dashboards tracking value realisation and conduct strict audits against "Shadow AI" data leakage. 6. Enable talent and cultural transformation: Oversee workforce upskilling and utilise IMDA guidelines for job redesign to ensure a human-centric transition. 7. Build customers’ trust: Utilise available frameworks and tools such as the AI Verify Testing Framework, which provides guidance on the responsible implementation of AI systems against 11 internationally recognised AI governance principles.

Section I: THE BOARD MANDATE

AI GUIDE FOR BOARDS IN SINGAPORE Section I: The Board Mandate

Articial intelligence (AI) has moved beyond the realm of future technology into the heart of corporate strategy. Understanding AI is no longer about technical mastery. For boards, it is about duciary literacy. Chapter 1: Why Boards Must Lead

Although the term “articial intelligence” was rst coined in 1955, the rapid advances and widespread adoption of AI as an operating reality in recent years have made it a critical boardroom priority.

1.1 A Primer for Directors

For boards, understanding e AI Lexicon is increasingly important. Dierent forms of AI carry dierent implications for strategy, workforce transformation, governance, cyber security, regulation and risk oversight.

“ Just imagine if your rm is not able to embed the tacit knowledge of the rm in a set of weights in a model that you control… you’re leaking enterprise value to some model company somewhere. ” Satya Nadella, CEO, Microso Singapore has elevated AI into a national imperative through its National AI Strategy. AI is increasingly framed not just as a business opportunity, but a critical pillar of national security, economic survival, social cohesion and future resilience. AI can no longer be treated as a peripheral technology issue delegated to management or technology teams. Boards must provide active leadership, ensure governance and develop strategic capability to facilitate informed decision- making.

AI literacy is becoming a core component of board eectiveness.

1.1.1 Leadership perspectives

e current “AI Spring” reects a fundamental shi in the relationship between humans and machines. It is not just another technology cycle or automation process. AI systems are redening value creation and reshaping how organisations function.

Consider these perspectives.

“I think AGI [articial general intelligence] will be the best tool humanity has yet created.” Sam Altman, CEO, OpenAI

AI GUIDE FOR BOARDS IN SINGAPORE Section I: The Board Mandate

The AI Lexicon

To lead eectively in the age of AI, boards must strengthen their AI literacy. At a minimum, directors should understand how AI is reshaping industries, operating models and competitive dynamics. e terminology continues to evolve rapidly, and many denitions overlap in practice. e descriptions below are intentionally simplied for the purpose of this Guide, and reect the landscape as of June 2026. • Articial intelligence: A broad term to describe systems capable of performing tasks typically associated

complete tasks on behalf of users, oen across multiple steps and systems. Unlike traditional AI tools which typically refer to rules-based algorithms, agentic AI can pursue goals, use external tools, make decisions within dened parameters and manage end-to-end workows, beyond isolated tasks. • Predictive AI: AI focused primarily on forecasting outcomes based on historical data. Unlike generative AI, predictive AI does not create new content, but instead supports areas such as demand forecasting, risk assessment and operational planning. • Multimodal AI: AI systems capable of processing and integrating multiple forms of data simultaneously, including text, images, audio, video and structured data. • Responsible AI: e governance principles, controls and practices designed to ensure AI systems are safe, ethical, transparent, fair and aligned with legal and societal expectations. Key concerns include safety, bias, explainability, privacy, cyber security and accountability. • AI Assistant: AI-powered tools embedded within workplace soware that assist users with tasks such as writing, research, analysis, coding and workow automation. ese systems are designed to augment human productivity rather than fully automate roles. • Autonomous system: AI-enabled systems capable of operating with minimal human intervention. Examples include autonomous vehicles, robotics, automated trading systems and self- managed operational platforms.

with human intelligence, such as understanding natural language,

recognising patterns, reasoning from data, and supporting decision-making. • Machine Learning: A subset of AI where systems learn patterns from datasets to make predictions or generate • Large Language Model (LLM): An advanced AI model trained on large amounts of text data that can understand and generate human-like language. LLMs underpin many generative AI tools and applications, including chatbots, copilots, search assistants and automated content generation. • Generative AI: AI systems that can outputs without being explicitly programmed for every scenario. generate new content, such as text, images, audio, video or soware code. Its business value lies in augmenting “knowledge-intensive” work, including summarisation, draing, research, analysis, coding and complex data synthesis. • Agentic AI: AI systems that can independently plan, execute and

AI GUIDE FOR BOARDS IN SINGAPORE

Section I: The Board Mandate

1.2 Key Shifts in the AI Industry

Active AI stewardship and board governance are critical to how organisations transform, shape value creation and manage emerging risks in an AI-driven environment. AI adoption is accelerating across every major sector of the economy. Gartner estimates that global AI spending will reach US$2.52 trillion (S$3.26 trillion) in 2026.

at phase is ending. 2026 marks a decisive shi from experimentation to execution.

Boards and executive teams are not just rushing into AI adoption to keep pace with hype. ey are now demanding measurable business outcomes rather than adoption metrics. Organisations are moving away from fragmented, employee-led experimentation towards enterprise-wide, strategy-led AI programmes focused on high-value use cases and operational impact. Along with greater investment discipline, the pressure for returns has increased. Investors and shareholders increasingly expect AI investments to demonstrate tangible business value within short timeframes. AI is becoming subject to the same governance expectations as any other major capital allocation decision.

e pace of change is outstripping traditional business and planning cycles.

By the end of 2026, 40 per cent of enterprise applications will incorporate task-specic AI agents, up from less than 5 per cent a year earlier. In healthcare, for example, AI is democratising access to medical expertise and diagnostics. In telecommunications, AI is modernising device management. Across professional services, manufacturing, logistics and nance, core workows are being automated, augmented and redesigned. AI disruption is changing the game, and the rules. Every organisation must assess how its business model, workforce, customer experience and competitive position is being reshaped and redened.

Boards should focus on measurable outcomes: productivity gains, revenue impact, cost eciency and competitive advantage.

1.2.2 From experimentation to scale

Boards must move beyond experimentation and oversight into strategic engagement.

One of the greatest challenges in enterprise AI is the gap between successful proofs-of- concept and scalable deployment. Many organisations can build pilots. Far fewer can industrialise AI across the enterprise. Leading organisations are therefore investing in the foundational capabilities required for sustainable scaling.

Boards must move beyond experimentation and oversight into strategic engagement.

1.2.1 e shi towards measurable value

e rst wave of enterprise AI adoption was characterised by the fear of missing out, or of being le behind.

Firms are establishing centralised AI hubs (“AI studios”) and enterprise platforms that

AI GUIDE FOR BOARDS IN SINGAPORE Section I: The Board Mandate

consolidate governance frameworks, reusable technology components, data infrastructure and specialist expertise. e objective is to move pilots into production within predictable timeframes. Strengthening data governance, cyber security, operational resilience, model oversight and vendor diversication enables organisations to accelerate AI deployment while reducing systemic risk. Sustainable AI transformation requires a shi from isolated pilot projects to integrated, platform-based operating models. 1.2.3 e regulatory environment As AI matures, legislators, regulators and standard bodies are making eorts to ensure that AI is developed and used in a reliable and safe way. ese eorts range from voluntary ethical frameworks to technical standards to enforceable obligations. ese can be at sectoral, national or global levels, and can also vary from jurisdiction to jurisdiction. Innovation is critical for companies' competitiveness, but boards need to ensure that governance requirements are taken into account in the process. Boards must therefore ensure that AI strategy is aligned not only with commercial objectives, but also with emerging legal, Boards must have the institutional capability to scale AI responsibly and competitively.

ethical and governance requirements.

• Global regulatory exposure: Singapore- based organisations may increasingly be aected by international frameworks such as the EU AI Act, particularly where they operate across jurisdictions or serve global customers. • Sector-specic accountability: In nancial services, the Monetary Authority of Singapore FEAT principles reinforce that boards and senior management are directly accountable for AI-related outcomes and risks. ese developments sit alongside increasingly complex obligations relating to personal data protection, cyber security, cross-border data governance and technology sanctions across geopolitical systems and global technology stacks. Governance frameworks and tools like the Model AI Governance Frameworks and AI Verify Testing Framework developed by the Infocomm Media Development Authority (IMDA) can help organisations navigate some of these obligations. ese are aligned with other international frameworks such as those from the EU, G7, OECD and the US, and help organisations to assess the responsible implementation of their AI systems. Responsible AI governance is a core component of enterprise risk management, regulatory, compliance and long-term organisational trust.

AI GUIDE FOR BOARDS IN SINGAPORE

Section I: The Board Mandate

1.3. Value Creation and Disruption

Boards must frame AI as a core business capability, integrated into strategy, operations, talent and governance. e Dual Lens of Governance is a balanced approach to Conformance and Performance .

expose the organisation to unacceptable legal, operational and reputational consequences.

Moving beyond “vanity metrics” such as chatbot usage or pilot counts, boards should focus on three core pillars of enterprise value: cost eciency, revenue acceleration, and risk resilience.

Focusing solely on risk may stie innovation; while focusing solely on opportunity may

The Dual Lens of Governance: Conformance and Performance

Singapore Institute of Directors (SID) advocates a balanced governance approach that addresses both risk oversight and long-term value creation.

Conformance (Risk and Governance)

Performance (Strategy and Value Creation)

Oversee AI ethics, governance and data management to prevent bias and ensure compliance with regulations such as the Personal Data Protection Act (PDPA).

Challenge management to move beyond cost reduction towards business model innovation and competitive dierentiation.

Manage AI-related operational, cyber security and resilience risks.

Identify opportunities where AI can amplify proprietary data, institutional knowledge and core workows. Allocate capital and resources to AI as a core pillar of long-term strategic capability rather than short-term experimentation.

Ensure AI systems do not reinforce bias or discrimination, reducing legal and reputational liabilities.

Establish accountability, oversight and assurance mechanisms for AI deployment.

Position AI as a driver of productivity, revenue growth and enterprise agility.

AI GUIDE FOR BOARDS IN SINGAPORE

Section I: The Board Mandate

1.3.1 Cost and revenue

In both public and private sectors, organisations are using AI-Enabled Productivity to redirect human capacity towards higher-value, judgement-intensive and relationship-driven work.

AI is being deployed to reduce repetitive administrative work, streamline operations and improve service delivery. Cost eciency and revenue acceleration require discipline in attributing business impact. Not all revenue growth associated with AI initiatives is necessarily caused by AI itself.

AI is a force multiplier for workforce productivity.

Boards should ensure management denes meaningful performance indicators that measure outcome rather than activity alone.

Clear measurement frameworks are essential.

AI-Enabled Productivity

Example: CPF Board

e CPF Board implemented an AI-enabled contact centre solution that signicantly improved operational eciency and citizen service outcomes.

Reported benets included: • Customer satisfaction: Improved rst-call resolution rate (90 per cent). • Time saving: Automated transcription and summarisation of calls saved 20,000 man- hours annually, freeing sta to focus on complex, high-empathy cases. • Scalability: System expansion time for peak periods (e.g., following policy announcements) was reduced from weeks to just three days. • Cost reduction: Operating cost per call lowered by up to 15 per cent.

Example: SingHealth

SingHealth implemented Note Buddy, a generative AI-powered clinical documentation solution that signicantly transformed clinical workows and enhanced the quality of doctor-patient interactions. Reported benets included: • Enhanced engagement: Allowed clinicians to provide full, undivided attention to patients and caregivers by reducing the need to manually type notes during clinic consultations. • Reduced cognitive load: Signicantly lowered administration burden on healthcare professionals by automatically transcribing and summarising complex clinical conversations in real time. • Multilingual eciency: Simultaneously captured and processed consultations across English, Mandarin (and Cantonese), Malay and Tamil. • Specialty customisation: Enabled clinicians to tailor AI prompts to their specic medical elds, ensuring highly accurate, structured and clinically relevant summaries.

AI GUIDE FOR BOARDS IN SINGAPORE

Section I: The Board Mandate

For example:

AI is also a topline growth driver, Scaling Expertise and Customer Value , particularly in sectors dependent on knowledge skills, advisory services and customer engagement. AI can help organisations scale expertise, accelerate onboarding, improve decision- making and enhance customer personalisation.

• Net productivity improvements may be more meaningful that hours saved. • Service quality may matter more than headcount reduction. • Customer outcomes may provide stronger indicators of long-term value creation than short-term eciency metrics.

Scaling Expertise and Customer Value

Example: OCBC Wealth Management

OCBC introduced a generative AI training programme for 900 wealth advisers that simulated real-time client scenarios and coaching interactions.

Reported benets included: • Performance gains: Advisers who completed the training doubled their weekly client appointments within the rst three months. • Increased client engagement: Clients reported more personalised engagement. • Revenue impact: Participants recorded a 50 per cent increase in revenue compared to the previous three months.

Example: CapitaLand Investment

CapitaLand Investment implemented a unied, data platform and integrated AI service to consolidate disparate data warehouses and accelerate AI adoption across its global business units. Reported benets included: • Cost savings: Shiing from on-premises data warehouses to a centralised cloud data platform saved more than S$1 million in operational costs. • Operational eciency: Automated processes and internal AI chatbots improved team eciency, saving more than 10,000 man-days per year. • Accelerated customer rewards: Leveraging AI and optical character recognition reduced CapitaStar receipt processing times from 3–7 days down to mere seconds. • Global data democratisation: Enabled real-time, 24/7 centralised data access for more than 4,000 internal users spread across 40 countries. • Enhanced customer experience: Launched “Cubby”, an AI-powered hospitality travel buddy for its lodging business (e Ascott Limited), giving guests real-time, tailored travel insights and itinerary planning.

AI GUIDE FOR BOARDS IN SINGAPORE

Section I: The Board Mandate

• Dynamic compliance management: AI- enabled compliance systems can monitor evolving regulatory requirements and map them against internal processes in real time, supporting compliance with frameworks such as the EU AI Act and Singapore’s Model AI Governance Framework.

Boards should challenge management to assess whether productivity gains are sustainably reinvested into growth, innovation and customer experience.

For example: • Deepen customer relationships. • Unlock adjacent markets. • Create new products and services. • Enhance pricing power and dierentiation.

Reducing operational and enterprise risk

AI’s ability to process large volumes of information rapidly creates new opportunities for operational resilience and enterprise-wide risk visibility. Examples include: • AI-enabled incident response: Agentic AI systems can assist with root cause analysis, log correlation and automated remediation during system outages or operational disruptions, reducing downtime and recovery time. • Predictive cyber security: AI systems can identify anomalous behaviour patterns (e.g., in network trac) that may indicate cyber threats before signicant breaches occur. • Supply chain and geopolitical intelligence: AI can support “horizon scanning” by analysing geopolitical developments, shipping disruption, market movements and external risk indicators that may aect business continuity.

1.3.2 Risk resilience

AI is reshaping enterprise risk management. As a risk management tool, it provides boards with the decision support to protect the company’s most important functions. Historically, risk oversight relied heavily on periodic reviews, static controls and retrospective analysis. AI increasingly enables real-time monitoring, predictive analysis and automated response capabilities. is enables boards to shi from reactive risk management towards active operational resilience, where systems do not just ag risks but actively work to neutralise them.

Strengthening compliance

Under SID’s governance framework, boards are responsible for ensuring that AI deployment aligns with legal, ethical and governance expectations. Examples include: • Algorithmic assurance and monitoring: AI tools can continuously test models for discriminatory patterns, unintended bias or governance breaches that could create regulatory or reputational liabilities.

1.3.3 Talent and human capital

AI transformation introduces signicant workforce and societal implications. Boards must ensure active Stewardship of People Resources , so AI deployment is human-centred and aligned with organisational trust and social cohesion.

AI GUIDE FOR BOARDS IN SINGAPORE

Section I: The Board Mandate

Key considerations include: • Preventing automation bias:

must oversee long-term workforce transformation and strategies that equip employees with evolving capabilities rather than allowing large-scale human capital obsolescence.

Organisations should maintain relevant “human-in-the-loop” decision-making frameworks for high-impact decisions, ensuring that accountability ultimately remains with people rather than algorithms. • Workforce resilience and upskilling: AI will transform jobs unevenly across industries and functions. Boards

Responsible AI adoption should strengthen workforce engagement and organisational trust.

Stewardship of People Resources

For Singapore, where human capital is the nation’s primary natural resource, AI strategy is inseparable from workforce strategy. • Upskilling: Job redesign and structured reskilling can help with workforce transition planning. AI is unlikely to replace human work entirely. Workforce capability planning, AI literacy initiatives, leadership development, structured reskilling and redeployment programmes will better position the organisation to transition to an AI-enabled economy. • Human-in-the-loop: Boards must ensure that humans remain accountable and there is ethical oversight in AI-supported processes. Building trust through human empathy and decision-making helps prevent automation bias and maintain social cohesion with customers, employees, partners, regulators and society. • Employee wellbeing: AI adoption can aect employee condence and workplace culture. Job anxiety and the risk of widening the gap between AI-enabled and non- enabled teams should be considered. Boards should oversee AI deployment in ways that maintain trust, inclusion, transparency and workforce wellbeing.

AI GUIDE FOR BOARDS IN SINGAPORE

Section I: The Board Mandate

1.4. Resilience in a Disrupted Economy

As a highly open economy, Singapore is particularly exposed to global technological shis. For instance, the rise of earlier technological shis like the internet and smartphones created entirely new market leaders, while disrupting incumbents that failed to adapt quickly.

Boards must ensure the organisation is agile enough to adopt emerging technologies while balancing governance, compliance and resilience with local regulatory and societal expectations.

Setting the tone from the top means ensuring AI is governed as both a strategic opportunity and a material enterprise risk.

AI is likely to trigger a similar wave of disruption.

Disruptors oen come from outside traditional industries. Boards must scan the horizon for non-traditional AI competitors.

AI GUIDE FOR BOARDS IN SINGAPORE

Section I: The Board Mandate

Chapter 2: The Case For (and Against) AI

AI is becoming a core driver of enterprise value, and moving beyond theoretical potential to real-world applications.

By examining both industry-specic and enterprise-wide use cases, directors can better evaluate how AI might redene the competitive landscape of their own organisations.

2.1 Defining Strategic Ambition

Boards should understand how AI can be leveraged to solve business problems: where value creation is already occurring, and how future waves of AI adoption will transform industries and business models.

Management may naturally gravitate toward safer productivity plays. Boards play an important role in ensuring the organisation does not focus solely on doing the same things faster.

2.1.1 Incremental to transformational change

Board directors must provide clear guidance and support for transformational change and doing things dierently.

Boards must help management distinguish between two fundamentally dierent approaches to AI adoption: Incremental Productivity vs Transformational Change .

Incremental Productivity vs Transformational Change

e “best-in-class” approach focuses on improving existing workows and operating models. AI is deployed to make current processes faster, cheaper, more scalable or more accurate. ese incremental gains are oen the easiest to implement and measure, and oen generate the most immediate return on investment. e “only-in-class” approach represents a fundamental shi in the operating model or value proposition. is is where AI transforms from an assistant to a system or an agent that executes repeatable, multi-step work. e focus could be on new business models, entering new markets, or “productising” internal intelligence.

Transformational change requires budget, resources and a healthy risk appetite. ere are also potential risks of distractions from the core business.

Ultimately, success may not necessarily belong to the rst mover. In many cases, the greatest value may come from being a disciplined and fast follower: learning from early market experimentation while scaling proven use cases eciently.

AI GUIDE FOR BOARDS IN SINGAPORE

Section I: The Board Mandate

2.2 Industry-Specific Perspective

• Education: rough tools like ChatGPT Edu, institutions like National University of Singapore are responsibly deploying AI across teaching, research, and campus administration to meet diverse academic needs. • Technology: AI becomes both the product and the driver of execution. Grab utilised AI for GrabMaps to improve mapping accuracy, and increased baseline accuracy from 67 to 80 per cent, signicantly reducing manual operations. • Small and medium-sized rms: AI is not just for large enterprises. Descript integrated AI-powered product

AI use cases are oen best understood through industry-specic problems.

• Financial services: AI is being utilised for high-stakes knowledge retrieval and research support. Morgan Stanley reported that over 98 per cent of its adviser teams use an internal AI assistant to access institutional knowledge, directly strengthening client-facing interactions. • Banking: AI is becoming an enterprise- wide capability rather than a narrow pilot. BBVA collaborated with OpenAI to deploy ChatGPT Enterprise to 100,000 employees globally, embedding AI into the bank’s fundamental operating model. • Retail and consumer businesses: e focus here is on scale and guest experience. Target uses AI for personalised recommendations and easier product discovery, showing how AI can drive value both internally and for the end consumer. • Healthcare and life sciences: AI helps scale high-quality care and reduce the administrative burden that oen plagues the sector. is is critical for boards focused on maintaining service quality and operational consistency.

capabilities, resulting in a 15 per cent increase in exports and a dramatic improvement in duration adherence.

Boards should assess how AI may aect the fundamental workows, economics and competitive dynamics within their respective sectors.

AI GUIDE FOR BOARDS IN SINGAPORE

Section I: The Board Mandate

2.3 Cross-Function Development

• Coding and soware development: is is a mainstream enterprise use case, helping teams write and refactor code. It is relevant for any business where technology delivery shapes competitiveness. • Workow automation and agents: is shis the focus from AI as an assistant to AI as a system that can execute repeatable work, leading to operating- model changes. • Data analysis and decision support: AI can help leaders interpret information and reason across multiple inputs to support strategic and operational work. • Multimodal understanding: AI is no longer limited to text; it can interpret images and real-world inputs. Grab’s mapping case shows how AI can interpret visual data in ways that matter operationally. • Personalisation and discovery: AI helps tailor experiences to ensure customers nd the right products or information more eciently.

Horizontal applications are the most broadly relevant because they cut across every sector. Boards should monitor these areas for immediate impact: • Employee productivity and work augmentation: is is one of the clearest starting points for any board. e value is in reducing friction across common tasks such as draing, analysis, and coordination. • Knowledge retrieval and synthesis: is is critical for organisations where expertise is distributed across many documents and systems. AI helps employees surface and apply this “hidden” knowledge in time- sensitive work. • Customer support and service operations: is area oers the most measurable impact. Klarna’s AI assistant handled 2.3 million conversations, representing two-thirds of its customer service chats, while reducing customer resolution times from 11 minutes to under 2 minutes.

AI GUIDE FOR BOARDS IN SINGAPORE

Section I: The Board Mandate

2.4 Emerging Disruptors

As AI technology landscape evolves rapidly boards should monitor the core capability shis occurring across the frontier AI landscape. A new generation of AI labs and infrastructure providers is expanding what technology can perceive and execute across both digital and physical domains. Some emerging capability areas include: • Spatial intelligence and 3D reasoning: Models capable of understanding, mapping and reasoning about three- dimensional environments. ese technologies underpin advancements in robotics, spatial computing, digital twins and complex physical simulations. • AI-native workow and soware engineering tools: Domain-specic AI environments that re-imagine how knowledge work and soware

development are executed, shiing the paradigm from basic task assistance to autonomous, end-to-end execution. • Specialised compute and hardware acceleration: Next-generation hardware architectures designed specically to train and serve massive foundation models at speeds and eciency levels beyond traditional silicon limits. • Advanced reasoning and strategic problem solving: AI architectures explicitly engineered to move beyond raw pattern recognition toward multi-step logic, handling complex, open-ended and unstructured strategic challenges. Individual companies will rise, merge, or disappear, but these underlying capability vector shis will continuously reshape competitive landscapes, talent requirements and operational capabilities.

AI GUIDE FOR BOARDS IN SINGAPORE

Section I: The Board Mandate

2.5 Potential Challenges

2.5.2 Technology and operational threats

Every new innovation inevitably has a double edge. Similarly, boards should keep an eye on the potential new challenges arising from AI.

AI failures can quickly become operational, legal, and reputational issues. Opaque decision-making, supply chain vulnerabilities, data leakage, and dependence on a small number of AI providers can undermine resilience and accountability. • Opaque architecture and explainability gaps: Frontier models and deep neural networks inherently lack mechanistic transparency, making it nearly impossible to trace exactly how an AI arrived at a specic decision. e “black box” dilemma creates severe regulatory and operational anxiety for corporate boards; you cannot govern what you cannot see. Without a clear audit trail, boards face immense liability regarding unmapped algorithmic bias, hallucinated data and compliance violations under emerging AI safety legislation. When an autonomous agent makes a catastrophic nancial or legal error, the lack of explainability leaves leadership completely unable to defend the decision-making process to regulators, shareholders, or courts. • Hyper-sophisticated engineering and supply chain vulnerabilities: Frontier models lower the barrier to entry for highly sophisticated cyber warfare. e risk shis heavily to the third-party supply chain and open-source soware libraries integrated into the enterprise architecture. An attack on a minor vendor can instantly compromise the parent rm.

2.5.1 Frontier threats

T he threat landscape has changed dramatically with the introduction of frontier AI models, the most advanced general- purpose AI models at any given time. Frontier AI models represent a new class of AI-powered vulnerability discovery systems. Testing by the UK AI Security Institute found that these models can operate without human guidance to execute highly complex, multistep cyber attacks. Frontier AI models can successfully map thousands of long- standing, unknown “zero-day” aws across major operating systems and browsers at machine speed. While export controls and safeguards are actively debated, the genie is out of the bottle. Attackers can now have access to “asymmetric oence”, where a single hacker or small syndicate can breach major corporate or government networks. Traditional corporate patch-management cycles that take weeks are completely obsolete when an AI-driven exploit can weaponise a aw in hours. Boards must shi from a compliance- focused security mindset to an active, machine-speed resilience strategy.

Page 1 Page 2 Page 3 Page 4 Page 5 Page 6 Page 7 Page 8 Page 9 Page 10 Page 11 Page 12 Page 13 Page 14 Page 15 Page 16 Page 17 Page 18 Page 19 Page 20 Page 21 Page 22 Page 23 Page 24 Page 25 Page 26 Page 27 Page 28 Page 29 Page 30 Page 31 Page 32 Page 33 Page 34 Page 35 Page 36 Page 37 Page 38 Page 39 Page 40 Page 41 Page 42 Page 43 Page 44 Page 45 Page 46 Page 47 Page 48 Page 49 Page 50 Page 51 Page 52 Page 53 Page 54 Page 55 Page 56 Page 57 Page 58 Page 59 Page 60 Page 61 Page 62 Page 63 Page 64 Page 65 Page 66 Page 67 Page 68 Page 69 Page 70 Page 71 Page 72 Page 73 Page 74 Page 75 Page 76 Page 77 Page 78 Page 79 Page 80 Page 81 Page 82 Page 83 Page 84 Page 85 Page 86 Page 87 Page 88 Page 89 Page 90 Page 91 Page 92 Page 93 Page 94 Page 95 Page 96 Page 97 Page 98 Page 99 Page 100 Page 101 Page 102 Page 103 Page 104 Page 105 Page 106 Page 107 Page 108 Page 109 Page 110 Page 111 Page 112 Page 113 Page 114 Page 115 Page 116

Made with FlippingBook Digital Proposal Creator