Daniel Huang at The Future of X(AI) Unconference

CEO explains how fashion ecommerce merchants are growing business with Rosetta.ai

Rosetta.ai CEO and co-founder Daniel Huang spoke at the Future of X(AI) Unconference.

Last month Daniel Huang gave a short talk about his Aesthetics and Desire AI product, and the importance of consumer insights for fashion ecommerce merchants. The talk was hosted by 500 Global and AI Yangjae Hub.

Hello everyone,

Nice to meet you guys, my name’s Daniel, co-founder and CEO of Rosetta.ai.

We focus our Aesthetic and Desire AI on fashion ecommerce, helping merchants increase their onsite conversion rate and order value.

So in this section I will explain how our applications help fashion ecommerce merchants gain deeper insights into their first-party data to boost business, and I’ll also go over the technology a little bit.

But first of all I will introduce our company. We were founded in 2016, at that time we were a project-based company but actually we were doing the same thing we do now.

Originally we made data analysis systems and recommendation engines to suit individual companies, but by 2018 we found some common demand so we launched our SaaS product, which became Rosetta.ai.

Then we became much more focused on fashion ecommerce and specialized in helping merchants extract the shopping preferences of each consumer, and predicting their shopping tastes.

With these consumer insights, our Aesthetics AI can provide personalized product recommendations that increase conversions and average order value.

In 2020 we got some press coverage in Forbes, making their Top 25 AI Startups list. We also won fourth place at RecSys, a global recommendation engine competition.

At that time the main focus of our technology was to extract the preferred attributes from the products.

And this is our core team. Actually we’ve grown quickly this year and even moved into a new office. Including our remote team members, we’re now a team of 32.

Currently, we help fashion ecommerce merchants double their average order value and triple their conversion rate.

Globally, we have over 1600 clients including Loreal cosmetics and Codibook, a Korean multi-brand fashion ecommerce site, who we help sell apparel products. I will share case studies with you later.

We provide consumer centric products and services through-out the consumer journey.​​

So what we do is focus on individual shopper’s preferences, so fashion ecommerce merchants can understand the preferences of each individual consumer and then provide personalized content to the shoppers.​​

We also help ecommerce retailers analyze their products. We use natural language processing and image recognition to extract the latent attributes from the product. For example, in apparel ecommerce, our Aesthetics AI can recognize the style, color, texture and many other attributes from the product image even if that data isn’t available. Image processing and object detection provide the attributes.

And meanwhile, we also collect the shopper’s onsite behavior. For example, when a shopper clicks a link of a product, we can predict the meaning of the click.

  • is it just a click?
  • or does it mean the shopper will select the item to add to the shopping cart?
  • or does it mean the shopper will purchase the product?

So with these insights we can understand the shopper’s deeper interest in the product.

Then we combine the preferred attributes and shopper’s behavior to build a preference profile for every shopper so we can match their tastes and their buyer intent.

So with our applications we provide four products for fashion ecommerce. Preference extraction gets shopper’s tastes so that fashion merchants can understand their shoppers’ preference distribution. This helps them prepare inventory for the next month and the next season.​​

For personalized product recommendations, our plugin supports various ecommerce platforms, such as Shopify, Woocommerce, and so on. The merchant just needs to paste a single line of code into their online store so their webpage will be able to provide a personalized online experience to their shoppers.​​

Our application lets fashion ecommerce merchants design personalized interactive content like pop-up recommendations and discount coupons for hesitant customers. When hesitant shoppers are activated, revenue and average order value increase. ​​

Our application works with communication channels like messaging bots, Facebook messenger, Line, WeChat and WhatsApp. It can also integrate with SMS and EDM. So personalized recommendations can cover every channel to detect the shopper’s interests and preferences and determine the right time to send the right offer to the shopper.​​

The current technology can help merchants label new product attributes even though the attributes are not shown in their product data.​​

And we also build up the preferences of every shopper. For example, our Aesthetic AI can predict that a certain shopper will like red more than green and stripes more than another pattern. Then…​

  • the red striped shirt is recommended to that particular at just the right time,​
  • or a discount pop-up for the red striped shirt can engage the shopper,​
  • or an email can be sent to remind the shopper to return to buy the red striped shirt.​

​So personalized product recommendations can appear on the Homepage, Product Details Page, Category Page, and most importantly, the Checkout Page. The context of the product recommendation is different for each page.​​

We provide several recommendation scenarios to help fashion ecommerce merchants choose which one they want to use to provide a unique, personalized experience for their shoppers.​

When customers go on a shopping website to browse around, they usually spend different amount of time on various products. Our company consolidates this data based on the length of time and the products to decide when and what kind of pop-up ads to show up​​

The shopper’s interaction with the product recommendation can be designed by the ecommerce merchant. We provide several interactive layouts like pop-up windows, sliding windows to trigger the hesitant shoppers to purchase more or make a decision.​​

​Personalized product recommendations can also be integrated into other channels like EDM, SMS, Chatbots and app notifications. So the shopper’s preferences are always at the center of the sales cycle, even at the end for cart abandonment.

When visitors add to cart but don’t check out the Aesthetics AI understands the shopper’s preferences and will automatically approach the shopper through a variety of channels to remind them to purchase the product.

And this is a case study of one of our customers, a multi-brand fashion ecommerce customer called, Codibook. They sell apparel products for young ladies in Korea, Japan, Taiwan and North America. ​​

They A/B tested our Aesthetics AI and after just one month, Codibook found our service helped them increase their revenue by 51 percent and average order value by over 38 percent.​​

Another critical difference Codibook realized was that our Aesthetics AI system optimized their workflow by reducing daily maintenance and providing better customer support.

Our technology is optimized for the fashion ecommerce vertical and includes apparel, cosmetics and accessories. ​​

So in each domain we can help merchants understand which product attributes are loved by which shopper and we actually help them build individual preference profiles for each shopper.​​

This is real personalization for each shopper, which is more advanced than many other solutions out there nowadays.

So we are looking for business and marketing partners to go global and we welcome any fashion ecommerce owners or merchants to contact us for more details.​​

We can help you to increase onsite conversion rate and order value and uncover the shopping preferences of all of your shoppers.​​

Ok, my name is Daniel from Rosetta.ai, Thank you guys.

About Rosetta.ai

Rosetta.ai offers fashion-optimized, image-based AI personalization for 1:1 onsite shopping experiences, authentic email marketing campaigns and preference analysis insights.

On average, our clients double their order value and triple their conversion rate because shoppers on their websites are more engaged by the industry-leading accuracy of our personalized recommendations.

Rosetta.ai has been featured in Forbes Top 25 ML startups and Analytics Insights Top 10 companies.

Sign up for a free 14-day trial today and start growing your business on day one.




Our mission is to empower merchants. Our technology lets merchants understand fashion-savvy customers and create unique shopping experiences that grow their businesses. 1000+ ecommerce shops using Rosetta.ai saw a 2x increase in AOV and 3.3x increase in conversion rate.

Recommended from Medium

Trending web technologies for 2019

Regulating ML/AI powered systems for Bias

Teach your conversational AI Chatbot to pick up entities in Teneo Studio

GetLinks launch Remote Income, a product to help millions of unemployed workers to access the…

AI in Online Casinos

Simplify Decision-Making with Holistic AI Synthesized Data

H+ Weekly — Issue #193

5 Language-oriented Tasks that Have Not Been Explored Enough by Deep Learning (1)

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store
Ian Mckinnon

Ian Mckinnon

Content Strategist @Rosetta.ai

More from Medium

Reflection Activity

Barilla SpA: A case study assessment

16 Experts Panelist Share Blockchain Trends for 2022

Future #61: Technology vs. technique