Shadow AI Is Already Inside Your Organisation
More than 90% of employees are already using personal AI tools for work tasks. The governance risk is not the AI. It is your data leaving your systems without any controls in place.
Chief AI & Innovation Officer
Terence Kok designs AI deployment frameworks for organisations that cannot afford to get it wrong — from national smart city infrastructure to first enterprise deployments. Governance, measurement, and accountability built in from day one.
Government & enterprise programmes →Engagements & Recognition · Press page →
What They Say
Before purchasing any tool, Terence runs a half-day diagnostic across five operational dimensions. It identifies where AI will return value, where it creates risk, and what order to proceed in.
Take the free self-assessment →AI Readiness Diagnostic — Five Core Dimensions
Half-day cohort diagnostic adapted from national infrastructure capability-assessment methodology. Leave with a scored readiness baseline and a twelve-week improvement roadmap.
Private advisory for organisations making first AI deployments — governance frameworks, vendor evaluation criteria, and measurement design. Scoped to your operational context and decision timeline.
Keynotes on AI governance, agentic systems, and measurable ROI for non-technical business leaders. Past forums include CDO Vision Singapore, InteracTech Asia 2026, and AI for Developing Countries Forum.
AI at Scale: From Pilot to Production
Enterprise AI deployment frameworks for leaders who cannot afford to get it wrong
A 23-chapter practitioner's guide from fifteen years of production deployments across Asia and the Middle East. Governance, measurement, and risk come first. Technology second. The eight-dimension framework in Appendix A is the foundation for every workshop and advisory engagement.
Every deployment produced a measurable outcome. Each SME Takeaway names the specific mechanism — not an analogy — that applies at any scale.
Domain-specific RAG platform deployed across six regional offices, trained on design standards, project documentation, and regulatory references. Governance model adopted as the Group template for all subsequent AI deployments.
A retrieval system trained on your own documentation consistently outperforms generic AI on your actual work. Classify every task: AI decides alone, requires a human check, or cannot act without approval.
Single governance standard satisfying Singapore's IMDA, Saudi Arabia's NDMO, and Oman's TRA simultaneously — without separate compliance stacks. Referenced in the Group's sustainability report as evidence of responsible AI practice.
Four questions determine your AI governance baseline: What can it do without asking you? Who checks the output? What happens when it is wrong? Can your staff still do this manually?
AI-first operational architecture designed before the airport opened — passenger flow, baggage, and airside management with no legacy systems to work around. Adopted as the reference model for future NEOM infrastructure builds.
Building the right measurement framework before you need the predictions is worth more than retrofitting AI onto data collected for other purposes. Design the data collection first.
Federated data architecture connecting traffic, utilities, emergency services, and civic services into a single operational view. Multi-agency coordination protocol adopted as the national standard for new city developments in Oman.
A single operational view of your business follows the same design as a city-scale control centre. What do you need to see, in what timeframe, and what triggers a human decision? Scale changes the volume — not the logic.
More than 90% of employees are already using personal AI tools for work tasks. The governance risk is not the AI. It is your data leaving your systems without any controls in place.
Standard SaaS contract templates were not designed for AI. Here are the ten clauses that actually protect you — and why most organisations only discover what's missing after something goes wrong.
MIT research is explicit: most AI pilot failures come from poor use case selection, not model quality. Here is the six-criterion framework for choosing right the first time.
A half-day cohort diagnostic. Each participant assesses their business across five readiness dimensions and leaves with a benchmarking scorecard and twelve-week roadmap.
What you leave with