Lalaland is changing the fashion game using AI

Providing a solution to fashion’s diversity problem

Online shopping is big business and according to Statistics Netherlands, Dutch consumers spent €1.9 billion in EU webshops in 2019. Buying clothes online is still a risky prospect, though. It can be hard for consumers to tell how a dress or top will look when they’re staring at photos of models who look nothing like them. Lalaland.ai, an Amsterdam-born startup, is offering an alternative in form of digital models of every size, age and race.

Co-founder Michael Musandu believes the young company can improve the online shopping experience, help brands embrace diversity and even reduce waste. StartupAmsterdam spoke to him about Lalaland’s journey and how it’s changing the industry.

How Lalaland came to be

According to Musandu, Lalaland's mission is simple: "We make photo-realistic, non-existing models for e-commerce web shops. This means we're generating non-existent people using neural networks to solve a diversity and inclusivity problem within fashion."

This is an issue that has plagued the fashion industry for years, both online and offline. "Have you shopped online? Before you purchase a product, right before you add it to the shopping cart, there's one major problem: you probably don't look at the model displayed in front of you. If looking from an ethnicity standpoint, the world is very diverse – we're not all one skin tone. If you look at the age of the model you're seeing, and the size of the model, it makes no sense to us that an XXL person should guess how the product is going to look on their body based on looking at someone who's a size small. It makes no sense to us that this is the way shopping is currently working.

I met my co-founder, Ugnius Rimsa, and it turned out we both had a passion for using neural networks to generate images through machine learning, and decided to immediately go to an incubator which works with universities, takes projects and tries to commercialise them and bring them to the market. As soon as we figured out that we wanted to be within the apparel world, we fell in love with this problem [the lack of diversity in models] and ever since then, we've surrounded ourselves with people in fashion to get to know about it more. It's taken us about a year or something to do real market intelligence and product development, and now, we're at the point of trying to take the industry by storm."

How digital models will change the e-commerce experience

Musandu: "The next time you're shopping, you can use the filters you usually have, like size small, medium, extra-large, and so on. Usually, these filters just show you if the product is out of stock, but [with Lalaland], you see the display model changing to better resemble you.

This way, you can see how the product will translate on your body. So, the biggest question we had is, 'Why aren't brands currently doing this?' It makes no sense that they are using perfect models and selling a dream instead of reality."

Why AI is good for brands

Musandu: "The costs involved with trying to shoot different-sizes models for different customer segments and different ethnicities 2 to 8 times a year, because collections come out rapidly, makes no sense from a cost perspective. It's way too expensive and time-consuming. That's why we came up with the idea of digitally generating a population of models who are very diverse and size inclusive.

Our vision for the future is individualising the shopping experience. So, you upload a selfie, put in some measurements – like your bust, weight and height – and you see yourself as an avatar. It's extremely realistic."

Not just inclusive, but reducing waste as well

Musandu: The biggest [advantage] is actually returns. The return rate for fashion items in the Netherlands is 44%, and this is 10% higher than the world average. So, it's great that we are in the Netherlands already, because the problem is very magnified here. It's about the environmental impact.

About 2.3 million tonnes of waste is generated from thrown-away returns. And there's unsold clothing – a lot of things get burned up in an incinerator or put in a landfill. So, the whole point is creating solutions to reduce the return rate. In theory, if you know what the product looks like on you when you buy it, you'll probably keep it instead of returning it."

Convincing companies to make the switch

Musandu: "There's a huge increase at the moment in digital fashion becoming a ‘thing’. As I said, the return rate is a big problem for brands, so they're looking for a solution. And we're coming from a cost-saving point of view, so it's very appealing for them to immediately try out the solution, because you're saving them so much. They also look at it in terms of conversion, because this can actually increase their revenue, and when we look at click-through rates, they are higher and people are more willing to interact with models that are displayed in different sizes and different ages and resemble them best."

Working with ACE Incubator

Musandu: "They say [working with ACE] is like a cheat sheet to the market and it's so true. There's a whole structure – they supply advisors, legal help and things like this, which helps with growth hacking a business. They also have lots of networking opportunities to meet investors if you need to raise capital and introductions to customer base and things like this, so I would say our success has come from working with an incubator."

Why Amsterdam just makes sense

Musandu: "Being based in Amsterdam was the best thing we ever did because of the ecosystem here. I can't thank Amsterdam enough for that. The ecosystem is extremely vibrant and there's always a lot going on. It's very much alive, no matter what stage you're in, so even if you're really in the startup phase, there's a lot of activity. Week in, week out, there are things happening and this is great in terms of resources and moving towards being a scale-up. There's no better city to be in. A lot of people think entrepreneurship only exists in Silicon Valley, but I beg to differ. This is an optimal location."