AI-driven master planning for industrial estate development

A leading developer of industrial estates in Thailand was hitting manual-design bottlenecks balancing roads, plots, utilities, and terrain – and each new estate required the full weight of senior planning expertise. Strict IEAT compliance rules meant that small errors invalidated entire layouts.

They needed to maximise saleable land while maintaining mandatory infrastructure and buffer zones, standardise design logic across planners, and translate planning concepts into precise CAD-ready engineering outputs – all without slowing delivery.

We built an AI-driven spatial drafting engine that generates zoning layouts, road networks, and plot subdivisions from CAD/GIS boundaries. A constraint-hierarchy engine applies the planning logic sequence – Boundary → Road Skeleton → Utilities → Plot Subdivision – so every draft respects IEAT rules by construction.

A topographic intelligence layer analyses terrain elevation to place drainage and utilities efficiently. The full system runs on-premise on NVIDIA RTX 4090 GPUs so no site data leaves the client, and a planned CAD integration layer will export AI layouts as AutoCAD DXF files with structured layers ready for the engineering team.

Tech stack: PyTorch, MLflow, Databricks, AWS.

Faster generation of industrial master-plan drafts with dramatically reduced manual workload for planners, and improved consistency with IEAT regulations across projects.

Secure AI-assisted planning runs entirely on-premise, giving the client a scalable foundation for AI-driven estate design across its future project pipeline.