Intelligent 3D Scenes in One Click

Spatial Intelligence’s journey began with a real-world problem...

SpaceForm Technologies - a leader in 3D world generation for architecture visualisation - faced a major bottleneck:

> 40% of their project time was spent manually selecting and positioning 3D assets - slowing delivery and constraining growth.

Scene composition - placing the right objects, in the right place, in the right way - was a critical pain point.

They needed intelligent automation to increase project throughput and scale their operations.

Over a six-month collaboration, we developed a single-prompt system that intelligently selected and positioned 3D assets - dramatically accelerating 3D scene composition.

The system combined:

1. Understanding of 3D assets (appearance, structure, function)

2. Spatial context (scene geometry and layout)

3. Creative intent (designer's high-level guidance)

4. Common-sense relationships between objects (e.g., chairs around tables, lamps beside sofas)

In short, the system made decisions a human designer would - intelligently selecting and placing assets in a 3D environment with minimal input.

Our [ solution architecture ← ] provides an in-depth technical breakdown.

1. One-click intelligent scene composition

2. Radically accelerated workflows

3. Early validation of market demand for 3D spatial intelligence

While the proof of concept demonstrated real-world value, it also exposed the limits of existing 3D AI models and techniques:

> Current vision-language AI models lack true spatial awareness and reasoning - the system could understand what objects looked like, but struggled to understand spaces and objects in 3D.

This project validated a clear commercial need - and revealed an even bigger opportunity:

> To unlock the next generation of 3D automation and robotics, we need a foundational AI model for end-to-end 3D understanding - not a patchwork of models trained for text or 2D images.

Spatial Intelligence is building that model:

We’re creating the core intelligence that understands the structure, content, and function of 3D environments - enabling machines to understand, construct, and interact with 3D spaces as humans do.

We seek to unlock applications across architecture, gaming and robotics with a single, generalisable model, without the need for retraining.

This is just the first step!