As consumers shoppers seek out new apparel in the wake of a global pandemic, fashion AI platform Savitude aims to pave the way for fashion design to become more inclusive by making fit possible across a wide range of body shapes and diverse populations instead of the hourglass body shape that’s been the industry standard that has driven design for decades. With retail sales predicted to increase between 6.5%-8.2% percent this year as the world begins to open up in the wake of COVID-19, according to the National Retail Federation (NRF), we asked Savitude’s CEO, Camilla Olson, and CTO, Nick Clayton, to let us in the company’s latest.
Tell Us About The Business, How Did It Come About?
Camilla: After graduating from college with a microbiology degree, I started out my career as an analyst at Johnson & Johnson. I worked for a number of years and eventually got my MBA. From there, I combined my knowledge in pharmacy/healthcare with my interest in business and took on a role as one of the very few women investors during the 1980’s. Next, after a series of investments for what is now the private equity firm Apax, I left investing and began to start companies of my own, first with corporate partners. After those successes, I founded three companies with ties to the pharmaceutical/healthcare space. Two of these companies relied heavily on predictive modelling, which from a thousand miles high would look similar to the process we use today at Savitude, albeit in an entirely different industry. The two predictive modelling companies ended up having successful exits.
And that’s where the story gets interesting. I had this whole background in business and science, but I was looking for the next thing. I happened to take my daughter on a college tour of a fashion design program that she was considering. I was fascinated. I wanted to be there. She didn’t end up applying, but I did. After years in the workforce and founding companies, I went back to school to learn fashion design and, if I’m being honest, to learn how to make a Chanel-type jacket, because I’d developed a bad habit there.
When I graduated, I continued working on the collection I started in the program and turned it into a luxury e-commerce label. As I was doing that, I learned firsthand about retailers’ and brands’ problems with returns, and how fit and, more specifically, shape with fit are a big part of the problem and what causes customers to return items. Suddenly, I saw a need in the fashion industry that I knew I was uniquely able to address with my predictive modelling experience. Combining my experience as a fashion designer with starting and exiting predictive modelling companies opened the door to starting Savitude, which uses artificial intelligence so fashion brands can create designs that better serve their customers, reduce returns and usher in a whole host of benefits ranging from lower carbon footprints to less waste and higher customer satisfaction.
Nick: Camilla reached out to me about creating a solution. I started working on the visual recognition system and knowledge base and we built that out and saw an 11% increase in sales with our first iteration testing with a major retail partner.
But at that time, there were about 200 other companies in the marketplace making recommendations, mostly on size, but still 200 different businesses trying to solve a similar problem, and for many body shapes there simply weren’t any clothes to recommend.
That’s when Camilla and I had another lightbulb moment, where we were both thinking we could take the technology one step further and actually design apparel rather than just recommending it.
So we took the leap to create what Savitude is today. It was a big pivot, but we wanted to solve the problem at the point of origin rather than patch the problem down the line, and the problem starts in the design studio.
So four patents later, with just a couple of designers and engineers, we built a knowledge-based recommendation system, a visual recognition system, including the human pose estimation, instant segmentation and classifier systems for robust image understanding, a drawing system capable of rendering novel designs and a generative algorithm that sits on top of all of these other systems and can create new designs while considering inspiration, trend, brand DNA, and of course body shape inclusivity.
More from Interviews
- Monica Eaton, Founder and CEO of Chargebacks911 and Women In Tech Judge, Explains What She’s Looking for From Entrants
- Meet Owen Thompson, CEO of Cambridge Future Tech
- Investor Insights: Everything You Need To Know About Index Ventures
- Investor Insights: Everything You Need To Know About Downing Capital Group
- Meet Sheryl Cuisia, CEO and Co-Founder, The Engagement Appeal
- Investor Insights: Everything You Need To Know About Sure Valley Ventures
- Meet Lenitha Bishop, Chief Operating Officer At The DPO Centre.
- Interview With Olly Craughan, Head of Sustainability at DPD
What Have You Learnt So Far?
Camilla: If you’re running a startup, having a sense of humour is really the best thing you can do. Think of it this way: if you take yourself too seriously, you will age much too quickly, and you’ll never live to succeed.
Nick: I could probably fill a few encyclopedias with the breadth of knowledge you need to wear all of the hats you need to at a startup, and I would be lying if I were to say I have learned all of it. The most important thing I have learned over the years is how to say “I don’t know that” and then how to find the people and the resources to help.
What Challenges Have You Overcome?
Camilla: One of the biggest challenges is educating people about the reasons for the problems with fit that so many women experience. Almost every day, we wow someone with the realisation that there are actually nine basic women’s body shapes — yet the fashion industry only designs and fits for one of them, the hourglass shape. It’s been the case for decades, and it’s no wonder so many women struggle to find clothing that not only fits, but also flatters their bodies. And it goes far beyond fit. For many women, clothing is at the source of a huge number of frustrations, disappointments and disempowering moments. So often women blame themselves, their bodies and turn an outdated industry practice into a commentary on their own appearance and, in far too many cases, their self-worth. Yes, our core business is licensing artificial intelligence to fashion brands. But our mission is really about helping women look their best. Nothing distracts us from that.
Nick: I think the most interesting challenge we have faced was the cause of our pivot. I remember trying to hunt down a bug that was causing our recommendation engine to return an empty list. I checked each step making sure nothing was going wrong, only to discover that there was in fact no bug. We had a system making recommendations, but simply no clothes to recommend to some body shapes. Taking our technology and reversing it to design clothes rather than recommend them required some creative leaps and a lot of tweaking, but we ended up with a much more robust system because of it.
How Have You Responded During Covid?
Camilla: We have always been a remote work-friendly company and have even won awards for our remote work culture, so the day-to-day for the team didn’t change. But during Covid, the need for a more inclusive fashion industry was underscored time and time again as diversity, equity and inclusion rose to the top of the wider cultural conversation.
Nick: We are proud to have created technology that allows for inclusive design that can serve diverse populations.
What Are Your Plans For Growth?
Camilla: Right now, we’re working on an exciting private project in AI design in which we’ve demonstrated that we can use artificial intelligence to not only design clothes, but translate those designs into virtual environments. It gives me great joy to know that we can do it.
Nick: Many large, household-name companies are trying to do this unsuccessfully. I know that the reason we’ve been able to do it is because of our focus on fashion design details and because we consciously chose not to use a brute-force algorithmic approach.
Learn more about Savitude at www.savitude.com and on Instagram @savitude.