Does the T-shirt go with those jeans? What do you think of the red sneakers? A few messages have already gone back and forward in the chat, when the next message pops up on the mobile. "Before we go any further I have to tell you something..." - "OK, what?" - "I’m not a real person."
Inspora is in fact a machine. Nevertheless, it is a likeable fashion fanatic and personal assistant – thanks to artificial intelligence (AI). Three students from Potsdam and Berlin developed the virtual fashion advisor, which finds out what users have in their wardrobe and what they like. Inspora draws on the outfits shown by well-known fashion influencers. "It asks, for instance, what do you usually wear to school or uni? And where do you go most frequently – to find out what the weather is like so that the outfits are always appropriate for the ambient weather," explains co-founder Willi Ibbeken.
The target group is Generation Z
A meeting at the premises of the start-up in Berlin’s Charlottenburg district – a ground-floor flat full of computers, monitors, beanbags and stands from their last exhibition at the back of a typical Berlin courtyard. The entrance is right next to a car repair shop. Is that a good omen? Many major IT careers began in garages. Inspora’s founders do at least seem to have hit a nerve – 50,000 users a month are already asking for advice, they tell us, and some four million messages are shared with Inspora. Most of them still come from the USA, because Inspora was launched in English in mid-2018. A German version has been available since the start of the year.
Inspora specialises in women’s fashion, and has a very young target group – most users are between 16 and 21. "They don’t want to call anyone and probably don’t have the money for a personal stylist," Ibbeken tells us. "But they are interested in fashion and they enjoy chatting."
Chat with an algorithm
Artificial intelligence is used to answer the young women’s fashion questions. And it starts with text recognition, explains computer scientist Daniel Birnstiel. "From the context of the chat messages we need to recognise what Inspora users expect." "What should I wear tomorrow?" thus becomes a mathematical challenge – and if Inspora, for instance, already knows that the user particularly likes the colour yellow or prefers wearing dresses, then it can finally produce the right fashion tip for work, uni, a date or a high school prom.
The second component of Inspora is what is termed the dialogue system. Conversations should feel real, and be varied. Inspora can bring together the items of clothing in users’ wardrobes and their personal taste with what is currently favoured by influencers and bloggers. The Inspora founders term this "matching" users and influencers. To this end the start-up works with several hundred of them.
Inspora learns from human stylists
The Inspora founders Willi Ibbeken, Daniel Birnstiel and Andreas Pavlenko, who studied together at the University of Potsdam’s Hasso Plattner Institute, essentially have users to thank for the Inspora concept. Originally they aimed to develop a shopping app. One feature was a chat, which was originally intended as part of the order process. But it was used more to discuss fashion matters per se, with up to 1,000 enquiries a day. For three months, the team answered every message themselves.
Then they began to automate the process, one step at a time. "We noticed that many of the fashion problems were very similar," Ibbeken tells us. The challenge was to translate the fashion tastes and interests of users and the creative styling process into specific data models with which Inspora can work. It took six months to develop the algorithm, which drew on the experience of over 100 human stylists.
Start-up promotion and prominent investors
Inspora attracted a prominent investor – Wendelin Wiedeking, who was CEO of Porsche for many years. He invested a high six-digit sum in the start-up and its self-learning chatbot. "Artificial intelligence is becoming increasingly important and will massively influence our everyday life in future," says Wiedeking. With Inspora he is investing in a solution that uses AI to give everyone access to their own personal styling consultant.
The start-up also had the support of the Federal Ministry for Economic Affairs. Ibbeken, Birnstiel and Pavlenko received a grant from the EXIST programme. "That was really important for us, to get the start-up up and running, build a team and develop a prototype," reports Ibbeken.
The EXIST grant programme covers the personal living expenses of start-up entrepreneurs from universities and non-university research facilities for a period of one year. For teams, materials and equipment costs of up to 30,000 euros are also covered along with coaching costs.
Inspora is intended to be entertaining
Users chat with Inspora via Facebook Messenger, but it is also accessible via its own website. The social media ecosystem is incredibly valuable for the founders, they tell us – Inspora users are generally already there. Generation Z spends up to three hours a day on messaging platforms, Ibbeken tells us.
He and his colleagues have given Inspora a personality so that it is fun chatting with her. She has pink hair, turquoise sunglasses and likes pizza. The start-up also offers individualised fashion news and make-up tutorials in the messenger, and aims to develop cooperation arrangements with fashion companies and publishing houses to this end.
The goal – a global platform
The founders have ambitious plans. Their goal is a global messaging platform for personal styling, fashion and entertainment.
Although Inspora uses AI to communicate with users, the team still answer some messages themselves. Like the e-mail they received from a young woman thanking them after Valentine’s Day. Inspora had provided inspiration for an outfit for a first date – which ended with a romantic kiss.