Functional AI For 3D Design Automation — From Path Finding To Generative Modeling For Building Construction

In the April 2026 CDFAM Computational Design Symposium, Richard—professor at Simon Fraser University and VP of AI at Canadian startup Augmenta—outlined a bold vision for functional AI in architecture, engineering, and construction (AEC). He highlighted that the construction sector dwarfs energy, healthcare, and entertainment in revenue yet lags far behind in AI adoption. This gap creates a lucrative opportunity for solutions that move beyond photorealistic renderings toward models that actually work: full interior layouts, structural integrity, and integrated mechanical, electrical, and plumbing (MEP) systems. By treating AI as a functional intelligence engine, Augmenta aims to turn text, image, or floor‑plan prompts into fully engineered, constructible 3D buildings.

Richard showcased several technical breakthroughs that underpin this vision. Early research on approximate pyramidal decomposition reduced 3D‑printing waste, while a Fermat‑spiral path‑filling algorithm demonstrated continuous toolpaths for complex 2D regions. More recently, Augmenta’s Visual Language Twin (VLT) bridges geometry with a domain‑specific language, enabling bidirectional, real‑time edits of building models. To overcome data scarcity and ownership constraints, the team synthesizes high‑fidelity LOD‑3+ building datasets, aligning spatial representations with large language and visual models. This synthetic approach provides controllable, copyright‑free training data, essential for scaling functional AI across diverse building typologies.

The business impact is already tangible. Augmenta’s proprietary MEP routing engine automated 16,000 miles of conduit for a Michigan elementary school, cutting engineer weeks of work down to days and integrating seamlessly with Autodesk Revit. Looking ahead, the company plans a foundation model for 3D construction that embeds functional knowledge—physics, motion, and interaction—into generative pipelines. For stakeholders in AEC, these advances promise faster project delivery, reduced errors, and new revenue streams from AI‑driven design services. Embracing functional AI now positions firms to lead the next wave of digital construction transformation.