AI GUIDE FOR BOARDS IN SINGAPORE
Section V: Organisational Resilience and Capability
For example, occasional employee prompting is very dierent from AI embedded into everyday workows, connected to trusted company context, and used consistently across functions.
Boards should consider which resilience gaps (across infrastructure, cyber security, data governance and operations) would materially constrain the organisation’s ability to scale safely and sustainably.
9.4. Ensuring Data Integrity
• Data accessibility: AI teams require timely access to governed data without creating security gaps. Self-service data platforms (e.g., Azure Synapse, Microso Fabric) balance agility and control. Data quality and integrity matter because AI systems are only as useful as the organisational context they can reliably access. Where enterprise information is fragmented, outdated, or di cult to retrieve, the quality of AI outputs and decision support will also be constrained. In many enterprises, the challenge is not lack of data, but lack of connected, usable and well-governed knowledge across systems. Clear controls over retention, access, encryption and data handling are part of a resilient AI data foundation, especially in regulated or sensitive operating environments. Boards should consider if the data foundations underpinning their organisation’s AI systems can withstand regulatory scrutiny, operational stress and adversarial testing.
Reliable, well-governed and accessible data is essential to trustworthy enterprise AI. Organisational value depends not just on having data, but on having usable, accurate and connected company knowledge that AI systems can draw on in a controlled way. e 2026 State of Data Integrity and AI Readiness Report by Precisely and Drexel LeBow identi es data governance as the pivotal factor for AI success. AI outcomes are directly inuenced by data quality and integrity. Poor-quality data produces unreliable outputs, while weak governance introduces legal, regulatory, operational and reputational risk. Boards should therefore oversee: • Data quality: Completeness, accuracy, consistency, timeliness and integrity of enterprise data used by AI systems. Automated data pro ling and cleansing pipelines reduce model errors. • Data governance: Clear ownership, de ned retention policies, role-based access controls, encryption standards, auditability and alignment with least-privilege principles. Regulatory obligations include Singapore’s Personal Data Protection Act and cross-border transfer requirements (e.g., APAC, EU).
87
Made with FlippingBook Digital Proposal Creator