How leverages AI to deliver magical online shopping experiences

Product updates   |   
Published April 30, 2019   |   

Shopping is inherently an intimate task. Whether you’re buying necessities, a gift for someone or you’re splurging on yourself. At the end-of-the-day, you’re spending hard-earned money on something you want to buy.
The shopkeepers of the past understood this. They would take the time to understand your needs, your likes and why you were buying something. And depending on these, they’d present the most suitable products for you. All the while, incorporating your active and passive feedback every time you shopped there.
However, small local stores slowly started turning into chain stores. And they in turn became global corporate brands. The concept of ‘personal touch’ was soon lost in a sea of commercialism. The coming of e-commerce only served to broaden the scale of shopping, both in terms of the number of customers and products. With this, the experience became more transactional, knocking the excitement right out of shopping.
Enter Crayon Data has teamed up with AI startup,, to help e-commerce companies recreate the magic for shoppers. With a little help from their friends, AI & machine learning.  ‘s algorithms and powerful AI engine are designed to get to know enterprise customers. It remembers their past interactions, understands their context, predicts the most relevant and personalized products. And then continues to learn and refine choices in real-time, by gauging customer reactions. Needless to say that all of this is done for millions of customers, across tens of thousands of products. Within 200 milliseconds.
So, from a business perspective these personalized experiences have led to an increase in repeat customers on e-commerce sites that have deployed As well as the purchase of products.

Taste based recommendations by powerful algorithms is the world’s only plug-and-play, omni-channel personalization engine for e-commerce portals. It leverages the power of AI, so that enterprises can recommend products from their websites that match a customer’s own preferences and taste. Thereby, driving a business benefit of a 8-13% increase in an enterprise’s top line.
The ever-learning algorithms find hidden patterns in an enterprise’s data. They identify the most relevant products for each customer based on their likelihood to click on a product or buy it. The algorithms are constantly learning and ‘genetically’ evolving to improve the predictions in real-time. Any new piece of data that’s made available to it – like a new click, a purchase, keywords searched or filters – is added to the database immediately.
The algorithms also use neural networks to enable real-time learning. New feature-sets are added regularly to cater to different use-cases.
They start with industry-standard collaborative-filtering and content-based recommendation algorithms, used with product embedding to make it scalable to large amounts of data.
Next, the context layer is added to cater to the nuances in choices. Based on seasons, date of month, day of the week, festivals, events, location and weather.
For example, would understand that the choice of apparel for a resident of Helsinki would be very different from that of a resident of Bangkok. It also keeps evolving with seasons, the weather and the latest fashion trends in the region. Moreover, the temporal layer is added to cater to recency of interactions, order of buying products, and frequency of purchases while shopping.

An omni-channel reach

Armed with the knowledge of the most relevant products for each user, the platform can display these products across every touch-point – on the website and app, with digital marketing and advertisements.
For instance, on every page of an e-commerce portal, EACH user is shown relevant products, reordered specifically to their likelihood to buy. Thus, increasing conversions by 8-13%.
Similarly, can personalize each advertisement (on Facebook, Instagram or the Google ad network), emails or notifications based on this customized order of products. Leading to an increase in click rates by 30-60%.

Ease of integration and use

One of the main differentiators of is the fact that it uses behavior and not identity to personalize experiences. The engine does not require a customer’s PII (Personally Identifiable Information) in order to recommend the best products, suited to their tastes.’s feature-rich application allows e-commerce portal managers to manage campaigns across channels from the interface itself. What’s more, with the portal managers can view and track the conversion funnel and drill downs in real-time.
An added advantage is that the team is geared to provide a smooth experience, while integrating the product with the business and its e-commerce portal. There are multiple integration options, including API’s, SDK’s in all languages, Magento Plugin and Google Tag Manager connector.
All this leads to clients seeing results within a week with!