What AI Startup Sarafan Wants You To Know About Native Advertising

Photo from Sarafan

Photo from Sarafan

By: Carli Evilsizer | July 11, 2019

We chatted with Ksenia Shirokova, COO of Sarafan Technology Inc, a startup that utilizes AI to identify objects in photos and videos then finds similar or identical items online for viewers to easily purchase. Ksenia shares more about Sarafan’s technology, native advertising success stories and the top 3 paid marketing tips every startup marketer needs to know.


Paid Marketing Tips From Sarafan COO Ksenia Shirokova

Sarafan logo.png



Sarafan Technology is a Marketing Tech company based in the USA and Europe. The company is a developer of interactive Advertising solutions for publishers. All the solutions are based on Artificial Intelligence technology Sarafan.AI that identifies objects in photos and videos published in media websites and searches for identical or similar items in online stores.



Ksenia Shirokova

COO of Sarafan Technology Inc


Ksenia Shirokova, COO of Sarafan Technology Inc., has 14 years of experience in fashion marketing and held senior roles at the likes of Next, Stockman and Dermosil Oy.

  1. So, what exactly does Sarafan do?

We’ve developed an Artificial Intelligence engine called Sarafan.AI which can identify products in photos and videos. Once recognized, our system seeks out equivalents in online stores and are presented in the form of a widget. Through this widget, users can automatically transfer to online stores and potentially continue shopping. Integrating our widget is extremely easy as most of the work is done automatically by Artificial Intelligence. As of today, our main client base is online platforms. We’re currently working with 122 media partners which include some big industry names, such as Glamour, Cosmopolitan and Elle in Russia, as well as Myself.de in Germany, Elle in Slovenia.

2. Who are your target customers?

On one side our target customers are media publishers such as fashion and lifestyle websites and for the other side we provide solutions for video content providers and platforms for video delivery such as streaming services and devices, video platforms.

3. What was the inspiration behind Sarafan?

At the start of 2010 artificial intelligence became a buzz word and people started wondering how it could work in a variety of industries. We decided to turn words into action and began developing a product which was based on computer vision and neural networking. By that point, neural networks were developed to the extent that they could recognize signatures, voices and faces. Product recognition at that point was still quite a niche, therefore, after years of development, in 2016 we officially launched our first product.

We developed an app to start with. As most of you will probably realize there are some fantastic platforms which allow you to save the looks, outfits and impressions that you found online. These features usually worked great at first, but viewers would quickly bore at the lack of further features. Therefore, our assumption was that customers usually have an “I see it, I like it, I want it, I got it” attitude. Therefore, instead of saving outfits, we decided to find them.

Photo from Tiffany’s  Twitter

Photo from Tiffany’s Twitter


4. What has been the hardest part about building an AI tech company?

As great as it is, but the problem arises with brand safety. In order to protect our advertiser's name, one of our key priorities was to make sure the artificial intelligence doesn’t associate it with negative images (for example those associated with death or violence). Textual recognition is by far the hardest aspect of developing an AI machine, as there are different languages and even within a particular language there are ambiguities. For example, you’ll intuitively differentiate the word like (as something likeable) and like (used to describe an example)- the AI won’t.

That’s why we hire special assessors who in cases of linguistic or contextual ambiguity can mechanically remove widgets from questionable content. The advertising partner can also equip us with “stop words” which we will later integrate when connecting his product to our image recognition.

5. How does Sarafan help companies improve their paid marketing?

With Sarafan technology, brands have an opportunity to create native advertising campaigns on online media and attract targeted traffic: those users who follow the link in the widget are interested in a particular product and are ready to buy it. There is no need to design separate creatives to attract audience attention, all photos and videos processed during the launch phase are ready for use.

6. What are some of the benefits of native advertising?

People are tired of online advertising: infinite teasers, obsessive banners are everywhere. The main advantage of native advertising is tact. It's not annoying. The second one is the ability to overcome “banner blindness”. And last but not least, native advertising matches with the user's interests and points at the spot.

myself 1.png

7. What are some common mistakes you see companies making with their paid marketing?

The most common mistake is the misunderstanding of the target audience. Marketers often don't know how to communicate with potential customers and where to find them. As a result, company places wrong ad messages in wrong places and waste ad budgets.

8. Can you share a few details about your most successful marketing campaign?

The marketing campaign of Guerlain new perfume on Cosmopolitan was the most successful. Sarafan AI’s technology was used to integrate the native ad to the Cosmopolitan website. The main objective was to present the product in a relatable context and lead the user to the online store for further shopping. In the course of the campaign, Sarafan’s widgets were placed in 338 articles on Cosmopolitan’s website. These included articles from the fashion, beauty, celebrities and news divisions of the magazine. The results exceed market averages by 13 times. ROI resulted 86%.

9. What is next for Sarafan in 2019?

We are ready to launch our new product – Sarafan Wishlist. Thanks to Sarafan Wishlist, users can save all their outfits in one “Favourite” cart and come back to it at any time. We make sure our system provides the user with a variety of purchasing options, up to date discounts and campaigns. Should the items the user chose run out of stock, we’ll ensure to deliver the most similar items elsewhere. So, the main goal for 2019 is to develop Sarafan Wishlist and provide the ideal service for users.

10. What are some of the benefits of Sarafan.AI for media publishers?

For media, Sarafan is an additional way to make the content more interactive, engaging and buyable. We offer our partners flexible revenue sharing business models where they can choose a monetization system which suits them best.


Sarafan’s Top 3 Paid Marketing Tips

1. Read Your Target Audience’s Mind

You should know exactly what your potential buyers love, what they live for, and what inspires them. You have to read minds to succeed.

2. Don’t Forget About Brand Safety

Buying digital advertising space automatically is very easy and effective, but not always safe. Remember about brand safety and make sure that your advertising does not appear in a negative context.

3. Do Tell Stories

We’ve loved stories since childhood. We are programmed to react to stories of a particular person than to facts and numbers. Therefore, the right storytelling to the right audience, is a key to marketing success.


It may sound cliché, but you’d be surprised by how many times brands miss the mark on their targeting efforts. It’s nice to see how precise Sarafan is willing to get to ensure their users are seeing content that’s not only right for them, but it's the stuff they’ll probably enjoy seeing AND engaging with!

Startup Video Marketing Tips From Swish, The New Video App for Entrepreneurs

5 Google Ad Word Mistakes Your Startup is Probably Making

3 Advertising Tips From An Ad SVP Turned Founder