Your stylist is a neural network
A group of young talented Belarusian developers has created a virtual stylist who can solve all the problems with the selection of a new wardrobe. We conclude our conversation with the co-founder of the project Kirill Sidorchuk and technology developer Vladislav Kaminsky about their innovative product Oyper.
– What is your uniqueness and what are your advantages?
K.S. Firstly, all our solutions are fully automated, they do not require any human involvement. Secondly, our smart recommendation engine is self-learning and monitors all changes in fashion trends. Thirdly, we offer customized solutions for retail. We have the technology, and we have developed 5 ways to use it, but we can implement any new ones and are very flexible in our approach to all offers and requests. I can give you an example from a request from one of our clients. We were asked to refine the classic cabinet based on our technologies, where there is a purchase history and a list of desired products. That is, to give the user the opportunity to constantly receive recommendations about what things are combined with the purchase history or with his . All similar solutions that exist on the market are not embedded in the personal account, they simply exist on the site (something similar is offered on the page of a particular thing). And we can equip every customer's office with our technology. This is a much more complex process. If the buyer has selected several items and they are in his personal account, then he is invited to choose items of clothing that match them in his personal account. In the next three months, the technology will be finalized, and it will be possible to pick up not only clothes, but also accessories, shoes and other wardrobe items.
Finally, another advantage − is the Camera AI. No one has similar proposals yet and, I think, will not appear in the near future.
– How does it all look from the consumer's point of view?
K.S. Well, for example, you go shopping and choose some thing for yourself. Let's say you found a blouse that you liked, but it costs 3 thousand rubles. As a rule, a mass-market buyer tries to find a lower price offer for a similar product. If your phone has a corresponding application, you can point the camera at the blouse and on the website of the online store to which this application is linked, see the entire range of price offers for the same blouse or similar products.
It turns out that your offer is primarily aimed at retail?
K.S. Yes, the B2B model was chosen by us for a reason. This is the direction where you can get a financial return the fastest. We are a startup and we are proud that today our technology is at the same level as companies that have received more than $ 10 million in investments, whereas we did everything ourselves, for our own money and in just a couple of years. Work on the idea began back in 2016, and the development of the technology itself began in March 2017. We were able to figure out how to achieve good success without a lot of funding. We began to receive some funding only when people saw what we were doing. Google supported us and allocated funds for the development of a small amount of $ 20,000, but with their help we were able to solve all the problems that we had. And now the company has given us access to its experimental platform, which will allow us to reach a qualitatively new level in the development of our solutions.
Have you analyzed the Russian market, who would you like to have as clients?
V.K. We are interested in all retailers in the field of fashion e-commerce. One of the main conditions is for the client's catalog to be large enough. Our model works effectively when there is something to choose from, there should be at least 5 thousand things in the catalog. At the same time, if the client uses all the solutions that we offer, his performance indicators can increase up to 30%.
– Do you already have clients in Russia?
K.S. In Russia, we cooperate with the company Getoutfit. She has a team of stylists who manually select images according to a questionnaire, that is, according to a questionnaire that the user fills out. And they asked us to automate these processes. Thanks to our work, they have saved time dozens of times. In addition, they began to offer users our functionality – selection with the help of a virtual stylist of things that fit the existing ones.
– Are you not interested in the B2C sector at all? Based on the above, you could well offer your services to stylists, who have now become quite a lot.
K.S. This is a very good question. The opportunity to cover the B2C market really exists. We recently realized that this is a separate layer of customers, which we didn't even pay attention to before. And we can really prepare an office for each stylist with very valuable tools that allow us to optimize work, save time, serve several clients at once and always be in touch with them. Now we are trying to solve this issue in Belarus. We have held a number of meetings with well-known Belarusian stylists to create a professional platform where the buyer can order a personal image that will be assembled by stylists and artificial intelligence. In addition, we set ourselves the task of developing a platform on which all the disparate Belarusian designers will be represented, not only already well-known, but also those who sell their products through social networks or in separate small stores. We want to give an opportunity to all young designers to develop their business, and to end consumers to get a quality product. If such a platform succeeds and it can be promoted in Belarus, then we will try to do the same in Russia.
– What will it be in price terms for retail, for stylists?
VKWe admit that at the moment we do not have a price list for stylists, we have not yet come to this. For retailers, we link the price to the number of items in the catalog. To date, the cost varies from $ 1,000 to $ 15,000 per month for a full set of functionality, depending on the number of items in the user's catalog. Moreover, it should be noted that the services provided do not affect the load and performance of the client's site in any way. All processing, all calculations take place on our side, and a team of three people is engaged in this, fully servicing the entire process, which we can scale to any number of clients.
We solve two main problems of the buyer. The first – he came to the store for a specific thing (things) and can't find it because of the huge selection and imperfection of the search engine. The second - the buyer wants to buy something, but does not know what exactly, and turns for advice and recommendation, expecting to get something that he would like. Objectively, the latter problem has not yet been solved by anyone, in any industry. If someone could offer the client exactly what he needs, it would be an ideal project. But we are moving towards this, and I think that after some time we will be able to create a similar solution.
Natalia Kulagina interviewed
Photo: provided by the developers