AI Design Workflows

Overview

I integrated AI image generation tools like Stable Diffusion and ComfyUI into B+H's design process to speed up early stage visualization. Instead of spending days setting up traditional rendering workflows, designers can now test facade options, lighting conditions, and material treatments in minutes. This lets teams iterate faster and have better design conversations earlier in the project.

The Challenge

Architectural visualization is time consuming and expensive. Traditional rendering workflows slow down design iteration, and communicating complex design intent to clients requires photorealistic renderings that take days to produce. The challenge: rapidly visualize large scale projects (entertainment venues, 50 storey hotels) from simple sketches while maintaining design control and exploring multiple facade treatments, lighting conditions, and material options.

The Solution

The workflow takes hand sketches or simple 3D massing models and generates photorealistic visualization options that clients can react to immediately. I built custom AI pipelines using ComfyUI for conceptual exploration and Automatic1111 for iterative refinement. ComfyUI gives more control through node based workflows for complex generation tasks. Automatic1111 is faster for quick iterations and inpainting to refine specific areas. The goal isn't to replace architectural expertise or traditional rendering. It's to compress the time between sketch and photorealistic image so design teams can explore more options and make better decisions faster.

Impact

Delivered client ready visualizations for complex projects including large entertainment venues and high rise hotels in a fraction of the traditional timeline. Enabled design conversations with multiple photorealistic options in minutes instead of days. This proved that AI can augment architectural expertise when used thoughtfully, compressing iteration time without sacrificing design intent or quality.

Technologies