Meinhardt AI Centre of Excellence
RAG platform deployed across 6,000+ engineers.
Domain-specific RAG platform deployed across six regional offices, trained on internal design standards, project documentation, and regulatory references. AI-assisted design review runs through human-in-the-loop governance gates, so an engineer, not the model, is accountable for every AI-assisted decision made on the platform.
Engineering teams across multiple regions were spending significant time on routine information retrieval, design review, and documentation cross-referencing. The challenge was deploying an AI system that engineers trusted, that met enterprise security requirements, and that improved measurable output without creating dependence.
Built a domain-specific RAG platform trained on Meinhardt's internal design standards, project documentation, and regulatory references. Human-in-the-loop review gates kept engineers accountable for AI-assisted decisions. Governance framework aligned with Singapore's IMDA MGF-Agentic guidelines.
Over 20% reduction in engineer time on routine tasks. AI-assisted design review deployed across six regional offices. Zero security incidents. The three-tier governance model (decide autonomously, human-checked, human-approved) was carried into the Group's AI deployment standards, replacing the ad hoc, project-by-project approach that came before it.
A retrieval system trained on your own documentation (product specs, proposals, service manuals) consistently outperforms a generic AI on your actual work. The governance rule is the same at any scale: classify every task by whether the AI decides alone, requires a human check, or cannot act without approval.
AI Task Classification Decision Matrix
The three-tier classification system used to govern every AI action on the platform: decisions the AI takes autonomously, decisions requiring a human review before execution, and decisions that require explicit human approval. Includes the classification logic, escalation criteria, and the audit trail structure for each tier.
Kok, T. (2026). Meinhardt AI Centre of Excellence [Case study]. terencekok.com. https://terencekok.com/projects/meinhardt-ai-centre-of-excellence/
The same diagnostic logic from deployments like this one, applied to your business in half a day.
Book the WorkshopDiscuss a similar programme- Meinhardt Group Annual Sustainability Report: referenced as evidence of responsible AI practice and governance maturity
- Singapore IMDA Model Governance Framework for Agentic AI: governance model aligned to published guidelines
- Employer testimonial letter from Meinhardt SUIT Pte Ltd (Managing Director, July 2025) naming this project available to procurement teams on request; not published online
Other case studies.
GovernanceGlobal AI Governance Framework
Governance standards for AI deployments across Asia and the Middle East.
Read case study →
InfrastructureNEOM Bay Airport Digital Strategy
AI-driven operational intelligence for a greenfield international airport.
Read case study →
City Master PlansSultan Haitham City, Oman
Smart city technology strategy and infrastructure engineering for a 1,570-hectare greenfield city.
Read case study →
Rolling this out across six regional offices meant six different cultures around what 'trusting the AI' even means, and the governance gates weren't red tape, they were what got senior engineers to use the system instead of working around it. I remember a reviewer in one office asking who takes the blame if the model is wrong; the honest answer, always, is the engineer who signed off. Zero security incidents isn't luck. It's what happens when accountability never moves off a person.
