Digamma.ai CEO Q&A Series: Jonas Cleveland, CEO of COSY
What transformative effects do you intend COSY to have on the retail sector?
We describe COSY as an aisle intelligence company that’s using machine vision and AI to improve retail execution and inventory productivity for our customers in retail stores and warehouses. The major trend occurring in retail today is the evolution of the retail store floor into a distribution center. This is something that COSY has been talking about for some time. For consumers, this means being able to order online, show up to the store and grab what you ordered plus other things, such as fresh produce.
For retailers, this means that they need more stores so that they can reach more people efficiently. Today we see stores like Target moving into more urban environments. We also see many stores closing down as the overhead of these stores is too high.
There is also the issue of an inefficient use of space. So, really, what COSY enables is the ability to optimize the way this real estate is used. Essentially, being able to optimize how you place departments, products and organize them on the store floor to drive revenue higher for retailers.
What made want to spin the company out of UPenn’s GRASP Lab?
To begin, I have to say that GRASP’s machine perception program is second to none. It has renowned professors and renowned research in this space. If you want to do something commercially, it’s a good place to be because The Wharton Business School is down the street as well. So, I went into graduate school knowing what my passions and my interests were but with no actual degree in mind because I knew that I simply wanted to do something entrepreneurially. So when the opportunity came up, this is what we decided to do, creating a company around the idea of using computer vision and artificial intelligence to make humans better.
How do you see AI and, more specifically, the field of computer vision changing the retail industry over the next decade?
In my view, every company is becoming a technology company or every company is becoming an artificial intelligence company. AI is pervasive in the way that it assists people in achieving their objectives and fosters understanding in data across an organization. In retail – the space where COSY plays – it’s about understanding whether a store is doing really well because of where it’s located or because the store manager there knows how to organize the store and run the store more efficiently. COSY enables retail companies at the headquarter level and at the corporate level to be able to understand the customers better, to be able to tailor marketing to the customers better, and to tailor experiences to their customers better as they walk into the store. So AI helps retailers be more efficient and also helps customers connect with products and with items that they are interested in.
We solved some major problems with computer vision, including things that COSY is now offering the marketplace. These include solving problems around fine grain classification, being able to detect and recognize all items in the view of the camera and being able to determine what the boundaries are around very difficult to recognize states.
What challenges did you and your team face when developing and commercializing COSY given that the technology was initially developed in an academic setting?
For someone who goes into retail, you don’t really choose to take on a market like this unless you see challenges as opportunities. So I think that we didn’t necessarily see the university as a challenge – we saw it as an opportunity. There’s no more cutting edge research that’s occurring than what’s occurring at the GRASP lab. Of course, there are always challenges in bringing a product to market but the university ecosystem was a partner as opposed to hindrance.
What new features will COSY have in the future?
What we enable a retailer to do is to open up the black box that is their store floor and allow them to understand what’s been shipped to the store and what’s leaving the store with a very high accuracy.
This is really powerful and, once you do that, you can enable all these additional applications from enterprise analytics to customer applications. So when you’re walking through the store, there’s an overlay of all the metadata and information. You can know using beacons where items are located in the store but you don’t really know how to pair that information correctly unless you have a computer vision platform that’s powering other devices or if you had a map of where all the assets are located beforehand.
With over a decade of AI experience, Digamma.ai’s team are your trusted machine learning consultants, partners, and engineers.