Bode and Microsoft’s AI-Powered Archive of Quilted Textiles

Dubbed the "Bode Vault," this system catalogs the finer details of an ever-growing collection

Emily Adams Bode‘s process favors the analog; sketching details passed along by collectors and artisans, repurposing vintage quilting textiles, enlisting skilled craftsman to replicate one-of-a-kind pieces and more instead of mass-producing her ideas. Her garments—released under her eponymous luxury brand, Bode—have won her the 2019 CFDA Award for Emerging Designer of the Year and the Inaugural Karl Lagerfeld Award for Innovation. As such, a collaborative project with Microsoft might seem out of character, but they worked diligently on a new system titled the Bode Vault, an AI-driven digital quilt expert that can identify pattern styles, eras, regions and more.

“I had wanted to begin preserving my personal archive ever since I started building it,” Bode tells CH. “But this came about through a conversation on how to take the next step in not only scaling my business but also preserving our incoming products and the product that I’m not actually utilizing or making clothing out of.”

The concept presents itself as an easy-to-use app that only requires a bit of focusing and a few taps on the screen. Using a tablet or a phone, Bode employees (the tool will remain internal for now) snap a picture of the quilted textile and a profile appears with its style, era, artist/maker, geographical location, fabric type, colors and other similar patterns. Trained on the Bode team’s comprehensive collection, the AI can identify patterns through complex color combinations and natural mistakes. This catalog also acts as a space to store the individual histories of each textile—the stories of generational craftsmanship and unwavering dedication to tradition.

“I wanted to not only preserve the narratives but also all the oral histories I was beginning to save by myself,” Bode says. She’d become reliant on tags hung from each swatch that summarized the piece in a few key details. Eventually, converting to a more sustainable system became necessary. “Also as a space for my staff to be able to learn about them, and a future goal is to have it be public-facing.”

The stories Bode had been collecting not only coincided with the specific fabrics she was trading, but also the larger history of quilting—a medium that, outside of museums, had remained largely free from technology’s encroaching. She wished for a tool that serviced the textiles and its keepers. She’d also previously drafted proposals for a similar tool, in a competition to get a grant, but had approached the idea more broadly—proposing the idea for a tool that would be used within a library of complex images.

“I was gaining a lot of information just by going to my dealers and learning. Then, when we go to put the product up online or when we go to sell it to our retailers or even direct to consumer like in our store, it was hard to tell the narrative and keep it all organized. We were putting it all on our garment tags. So, this is sort of the next step,” she says. “I don’t even really use a computer myself. I don’t think I foresaw building it out in this way, but it was a really natural solution. Thinking of what already exists for image recognition, this is a natural next step for us. Quilts really were the best place to start because they can sometimes have a recognizable geometric pattern.”

Images courtesy of Microsoft and Bode