Digitized design process accelerates response to fashion trends

By Hildegard Suntinger

Anna Franziska Michel was a fashion designer herself and experienced how time-consuming and inefficient the design process can be. Considering the rapid change in trends in fashion, it’s clear that drawn-out development processes must inevitably lead to backlogs. Data can change that, she thought, and began researching AI modules for digital product development with a research group at HTW Berlin’s Department of Business Informatics. Now Yoona.ai, her software-as-a-service solution, is running partially automated and has already been tested in large German fashion companies. In this installment of the Start-up-of-the-Day series, Anna Franziska talks about her innovation and the challenges of starting up a company:

What problem are you solving and why is it important?

Before fashion designers even know what to design, they have to gather information, such as the current bestseller in a category. Artificial intelligence manages this laborious research in just a few seconds. To digitize the finished design, conventional programs first have to make a 2D cut to create a three-dimensional model.

Currently, only 10 percent of all fashion companies can create digital or 3D products without first making a 2D cut. We want to make this technology available to everyone. Our software should make it possible to create a 3D object with just a few clicks and then use it for multiple media channels. And the faster fashion companies can respond to current trends and the needs of end consumers, the more likely it is that their products will be sold on the market and surpluses can be avoided.

How exactly will the design process be automated?

Users of our software can create a profile on our platform and upload basic things for the product development process such as color data, mood boards and technical drawings. In the partially automated design process, artificial intelligence in the form of two neural networks runs in the background. One of the networks learns using the data – product images, web analyses, color analyses, etc. – and the other designs digitally. Both neural networks must be trained to the specific requirements. Once the training is complete, they create up to 20,000 design suggestions in a matter of seconds.

What was the biggest obstacle you had to overcome?

Before founding my start-up, I first had to work with a research team at HTW Berlin to find neural networks that are applicable in the field.

It’s also a challenge to find the right team – to decide who fits in and understands the product. For a programmer or a normal technician, a T-shirt is a square with a hole in it. But for a designer, it’s different. A technician needs to have an understanding of what a designer does. My co-founder, for example, is very good at that.

What have been the best moments so far?

There were great moments, such as when we achieved the shortening of the process and when we realized that our product really delivers the expected impact. It was also great when companies started using our product and realized the added value. Later, when we won awards, we were proud and happy because it’s nice to get something back for all the work you have been doing.

But above all, it wasn’t easy to get involved in a subject that involved highly complicated technology since I was actually a fashion designer. Just taking this step from designer to software company was a big success for me.

How difficult was it to get funding?

As a woman? (laughs) I think it’s always a challenge to get funding. But an additional hurdle was definitely that our innovation is anchored in fashion and that I, as the founder, am a woman. But we did it and closed a very normal pre-seed round. Techstars is an American accelerator that invested in our start-up, as did Alexander Gedat, the former CEO of German fashion companies like Gerry Weber and Marc O’Polo.

What are the conditions like for your start-up in Berlin?

I previously lived in Madrid and considered staying there. Today I have two children and a start-up and probably wouldn’t have been able to balance that anywhere else. For me, Yoona.ai is not just my start-up and the eight people who work there, but a network that consists of some 100 people – and most of them live in Berlin. We have supporters like the University of Applied Sciences (HTW) and the city of Berlin. Our partners include Premium Exhibitions, one of the big German fashion trade fairs, where we are co-curator in the Fashion Tech area. That’s our network and relationships from Berlin as a location.

Where would you like to be with your company in five years?

Our design process is already partially automated but we will constantly continue to research to develop it further. We also just founded the conference series yoonaverse and want to build The Metaverse Berlin together with partners like Telekom, H&M and some Italian brands.

Our goal is to become a global company and THE provider in digital generation of apparel. I’m flying to New York next week for the PI Apparel conference and then we want to start rolling up the American market.

We want to become the Tesla of the fashion industry.

Tesla in the sense of sustainability?

Yes. It’s crucial that people understand that technology can have an impact on the sustainability of products. With technical support, you can calculate much better what is needed to design and are faster and closer to the trend as a result. The 3D rendering of models can reduce the need for physical prototyping. Digital models make physical samples obsolete for prototyping, photo shoots and fashion shows – making the design process more sustainable.

With my start-up, I wanted to make working with new technologies more feasible. As a designer, I would have been just one of many, but with Yoona.ai I can change the back end of the whole industry and create an exponential impact.

Be sure to read other instalments in this series.

Each day, Innovation Origins puts a start-up in the spotlight. Read them all here.