Meet visionPigeon: the Visual AI Shopping Advisor

Today it is our pleasure to unveil the 2nd startup (out of 10 selected so far) from the StartupYard Remote Lab program. visionPigeon is a computer vision company specialised in helping fashion eshops and their customers with better product inspirations. We sat down with Jan Kotek, the CEO and co-founder of visionPigeon to talk more about what they do and their plans for the future.

Hi Jan, tell us a bit about your personal journey towards founding visionPigeon. How did you get here?

It all started when I met my co-founder Radim, our CTO and the creator of our visual AI engine. He was telling me about what computer vision combined with machine learning can do and from that moment I was thinking what value it could have for e-commerce. 

Jan and Radim, CEO and CTO of visionPigeon

It is in e-commerce and B2B where I feel strongest. For over 5 years I have been creating digital products for B2B and for the last 3 years I have worked at Heureka, where I was responsible for developing our business development activities for Czech-Slovak e-shops.

But my experience of building products is, first get to know the customer’s problems well, validate the interest in the solution and then you can start creating.

Before we had anything created, we approached our first client and I showed them in a Google slide presentation what we could do for them. They liked it immediately so we started creating the first product.

What is visionPigeon for and what problem are you solving?

VisionPigeon increases revenue for e-shops through advanced product inspiration based on visual artificial intelligence.

For e-shops selling fashion or furniture, we solve the problem of low conversion rate, i.e. a user comes to the e-shop, doesn’t choose anything and leaves. In fact, the e-shops pay a significant amount of marketing money for each visit. 

Another problem is the low number of items in the shop cart or bad customer retention.

You are not the first startup to try solving this problem and there are already available proven solutions. Why should a company look at visionPigeon?

Compared to other solutions, our model works primarily with the visual aspect of the product (shape, colors, design, pattern,…) when evaluating the recommendation, so we get many times more information about the product than we would get from the parameters or description. This allows us to select the right product much more accurately.

Zoot – similar products powered by visionPigeon

How big is the team and why are they a good match for this project?

We have grown very quickly from three founding members to the current seven people. We have been working together for almost a year now, selecting only A players based on recommendations. Each new member has to go through a round of interviews with each of us and everyone has to approve them. This verifies both the candidate’s knowledge and cultural fit.

What have been your team’s biggest personal or professional challenges in making this project a reality?

The biggest question for me was how the team would work remotely. We have been building the company remotely since the beginning, there was no other way. We were founded in the “covid” time in September 2020, when we were all at home and in addition, the team members are sitting in Prague, Brno, Liberec and Pilsen.

The advantage is that we can recruit talent from different parts of the world and anyone can decide to work from the beach in Bali, for example. The disadvantage is that we rarely get everyone together, so we organize weekend off-sites every quarter. 

What did you see as fundamentally wrong with the fashion ecommerce industry?

The biggest problem is the lack of technology. E-shops would like to recommend better, which would positively affect their revenue, but they don’t have the tools to do so.

That’s why it’s common to see an e-shop recommending a bag for your high heels or a women’s bra for a men’s bathrobe, for example. The products in the Most Similar section are not similar and you will rarely come across an e-shop recommending suitable accessories for what you are currently ordering, which is a knock to the shopping experience and a missed opportunity for the e-shop.

How does visionPigeon help solve that issue?

At the core of the solution is a visual AI engine that analyses the appearance of products. To do this, we process information about the products from the parameters and last but not least, we have information about the user’s behaviour in the e-shop. We can see, for example, what products they view, buy or add to their favourites.

Let’s talk more about your products. What would you say distinguishes you from other similar solutions? What is your “killer app?”

Our visual AI platform offers very advanced forms of product inspiration for users. The Most Similar module is very popular, where users see the most similar alternatives on a product detail. Which comes in handy when you don’t like one small detail on the product you’re viewing, you have similar alternatives right at hand. 

The killer app is our solution to “Shop the Look”. This allows the e-shop to offer customers to purchase an entire outfit from almost any product photo. Our visual neuron automatically detects the products that appear in the photo. This inspires the customer to see what can be purchased with the product, increasing the chances of the user taking home more items.

visionPigeon instantly detects all products on an image allowing customers to buy individual items or the whole look

So if you are an eshop owner in the fashion industry, what do you need to get started with visionPigeon? 

In the first phase, we talk to clients about what key metrics they want to improve in their e-shop. Depending on the goals, we choose a solution and a way to implement it. We always start with a three-month pilot to let the e-shop test the benefits of the solution.

In terms of implementation, for e-shops that don’t have development or capacity, we can do everything necessary for the launch on our side. What we need from the client is the insertion of a few lines of code into the page and a product feed from Heureka or Google Shopping.

For other e-shops, we have an API that we use to communicate with the e-shop system and tell it what products to display on a specific site for a specific user.

Is visionPigeon available now? 

Yes, we went live very quickly. We visited our first client in September and by December we were already helping Zoot sell more. We were able to increase traffic to the Most Similar Products section on Zoot by 78%. 

By the way, you can find more at

How many clients do you have today and in which countries are you seeing the biggest interest?

Besides Zoot, people can find us for example on T.M. Lewin and we are in discussion with a lot of the biggest fashion e-shops in the Czech Republic.

Our solution is for bigger fashion and furniture e-shops, so it is necessary for us to move quickly from the Czech Republic to other developed markets. We consider the Czech Republic and Slovakia as our home market. Subsequently we want to enter Germany, France and then the UK. 

What do you hope visionPigeon is going to be in 5 years, as a business, or as a technology? What would be your ideal scenario?

We will do everything possible to have the company in all major markets in Europe and the United States at that time. We will be helping hundreds of customers and we will not stop with just personalised recommendations.

What do you hope that visionPigeon will be able to do for its customers in the next few years that they can’t do now?

VisionPigeon always looks at the needs of our clients. I can imagine us venturing into the area of fitting or creating personalised content marketing. The goal is a visual AI platform that helps e-shops bring in customers and provide those customers with the perfect product selection.

Talk a bit about your experience with StartupYard. What were the surprises, challenges, highs and lows?

We are still at the beginning of the programme, we still have a fundraising phase to go. I don’t have any complaints, I’m just sorry that we couldn’t physically meet with the other teams and work on projects together in one co-working space.

I can say that participating in the program has moved us forward in a very meaningful way. It would have been hard to reach dozens of mentors and potential investors in such a short time. Some of the mentors have opened doors for us to potential clients. I find invaluable the practical advices from Cedric and Nicola, with whom I have calls on a regular basis.