Digamma.ai CEO Q&A: Amar Chokhawala, CEO of Reflektion

Leave a comment
Artificial Intelligence / deep learning / Machine Learning

Digamma.ai CEO Q&A Series: Amar Chokhawala, CEO of Reflektion

Reflektion seeks to give retailers a deeper understanding of customer intent. How does your company do this from a technology perspective?

In today’s world, everyone is sharing everything. So, the customer’s intent almost starts with temporal signals. When I’m looking for a place to eat in the morning, that means I’m looking mostly at restaurants which are actually serving a breakfast or lunch. Here’s an example of how the temporal aspect comes into play. Imagine a three year old—if you asked her what an “apple” is, she would say it was a fruit. If you posed that same question to a thirteen year old, she would say that it’s the technology company. So, in order to understand customer intent in retail, we need to build a comprehensive user profile that takes into account this temporal dimension. Because if you don’t know the consumer, then you won’t know what the intent of that particular consumer is. Our technology is capable of building those user profiles. We focus on the user first and, to date, have built 600 million plus profiles. That allows us to gain a deeper understanding of customer intent. We use machine learning because, unlike other technology, it works more effectively with more data. The more data you have, the better the model performs.

Your company focuses on responsive merchandising—why is it important?

I think the way retail always has worked, with the United States taking a leading role and the rest of the world adapting to its example, is through generating a demand. There are two main ways retailers generate demand: the first way is that you come up with a very cool innovative product, whether it’s a jacket, a pair of shoes, a bag, etc., and you create a brand around that product. The second way is through promotions: you continuously tell users that you have a 40% off, 50% off, 60% off sale this weekend. Think about it—everyone you know does a Black Friday promotion. Consumer behaviour is actually conditioned to give that promotional strategy a response.

The question we ask is: what happens when consumers come to your website or any sales channel and they show specific intent? For us, a responsive merchandising strategy is all about how you respond to that intent. So, for example, if I’m looking for a bag and I show intent that I’m going after a high-end bag, will my merchandising strategy, in the user’s same session, change to show me high-end brands? You can complicate this example even further by communicating that you’re looking for high-end Prada bags that are made out of a specific leather and within a certain price range. Even further than that, a given consumer profile within an effective responsive merchandising system will not just include information on that individual consumer, but also information on that consumer’s spouse. If I was shopping for my wife or if my wife was shopping for me, that shopping sessions are completely different. So the responsive merchandising strategy needs to be completely different in both cases. Responsive merchandising therefore always has to be tied back to that central concept of a user profile.

How do you see machine learning changing the retail industry over the next few years?

Machine learning will fundamentally shift from product lifecycle management to supply chain merchandising planning, all the way to visual purchasing, email marketing, and retargeting. All retailers look for an item that they can source from the U.S., China, India, etc., and this can happen within a range of three to twelve months. Retailers must then decide what quantity to buy based on a prediction of how well the item will sell. And then they have to figure out how many they’re going to sell within the first eight to ten weeks. If they sell through everything within that period, they find out the hard way that they underestimated the demand for the item and did not order enough. And conversely, if they sell through very little, they will have to liquidate the rest of their stock. So, they have to strike a perfect balance. Machine learning is perfect for retail because it has the ability to predict and understand consumer preferences—it will become the foundation of retail.

Why did you choose retail as a sector to go into?

Our main thought process of selecting retail was simple. Retail is the most responsive sector in terms of finding out whether your technology will work or not. A single session for a user lasts, let’s say, 30 minutes. If you want to prove your AI or machine learning technology works, it is much easier to demonstrate its real business value when the transactions are happening within that short period. Compared to other sectors where transactions may happen over a longer period, such as travel, those purchasing decisions are happening over several weeks due to the consumer’s extensive research. So, it takes longer to prove the benefits of the technology.

What are the next set of projects or new features that you are looking to add to your platform?

So, we are heavily investing in a deep learning algorithm, where we are moving away from statistical models like regression, to an algorithm that has multiple layers of neural networks. We feel that this algorithm will provide much better results. We are also creating long- and short-term memory models that are capable of doing temporal-based learning.

With over a decade of AI experience, Digamma.ai’s team are your trusted machine learning consultants, partners, and engineers.

Leave a Reply

Your email address will not be published. Required fields are marked *