Fashion design teams spend huge amounts of time moving between disconnected tools, from sketching and moodboarding to creating tech packs for production. Most of this process is manual, repetitive, and hard to track, leading to lost time, creative burnout, and delayed launches.

To solve this, we built Artifax, a unified platform that connects research, visualization, and production into a seamless workflow. This case study focuses on the 0-to-1 design of its flagship feature and the core of the ecosystem - Dreamboard, the AI-native canvas.

I led the product and design journey for Dreamboard, from foundational research and defining the core AI strategy to shipping the end-to-end canvas experience.

This case study explores how our AI native canvas Dreamboard improves the way designers research, create, and refine their work.

Our journey began with an observation from our existing B2B product Catalogix, a platform for managing and publishing digital product catalogs. Catalogix worked well, but it only solved the final part of the workflow - commerce. Through customer talks, one thing kept coming up. The real struggle was the middle of the process. Turning rough ideas into a finished design took time, money, and a lot of manual effort.

Research into this pre-production workflow revealed that existing design tools (e.g., Illustrator, Figma, Miro) were fundamentally unintelligent digital corkboards. They forced highly-paid creative professionals to spend up to 80% of their time on low-value manual labor, such as resizing, alignment, and layout revisions, instead of focusing on core creative and strategic design work. This established a clear opportunity to expand our value proposition by automating this inefficiency.

To move beyond assumptions, I initiated a deep discovery phase to understand the nuanced realities of the fashion design workflow.

Research surfaced four recurring pain points across design teams. Each one reveals how tools, feedback, and context-switching sap creative momentum.

How might we consolidate the fragmented 5-tool workflow (research, ideation, collaboration, production) into a single, intelligent canvas that automates the 80% of manual, repetitive work (like layouts, revisions, and data entry) and gives designers their creative time back?

After defining the problem and validating our hypotheses, we used a structured framework to realign on user needs. The takeaway was powerful, users didn’t want another whiteboarding tool. They wanted an intelligent teammate by their side.

Dreamboard brings two sides together in one place.

First, it gives designers every advanced whiteboard tool they expect. You get artboards, vector tools, text controls, a precise grid system, and strong alignment features. This set brings the same quality level as top whiteboarding tools, so switching never feels like a downgrade.

The second side is where Dreamboard gets its real strength. It connects directly to the brand’s ecosystem, giving the canvas clear context, memory, and direction throughout the creative process. This context comes from Scout, Projects, and Knowledge, which together help the AI understand what the team is trying to create.

Dreamer is the creative copilot inside Dreamboard. It reads the context from Knowledge and Scout, understands what the project needs, and turns simple instructions into fast, editable output. Designers pick the direction and Dreamer builds the starting point, removing the slow setup and repeat work that normally blocks creative flow.

From here, Dreamer expands into several focused tools that support the full workflow, including AI board creation, quick edits and variations, resizing, and video outputs.

This is our STAR feature that replaces routine revision work. Before this feature, users were still spending hours doing small edits. People marked notes on top of images to adjust colors, resize logos, swap photos, or tweak layouts. Another person then had to recreate each fix manually. This slowed every project and created long review cycles.

And now Cohost handles the edit in seconds. Tasks that used to take an hour across multiple tools now take less than five minutes. This keeps designers focused on choices, not cleanup.