The line at the grocery store can be quite daunting, especially if you are one of those customers who has one item to bill, and the person in front of you has two shopping cart full of products. It was something that Sanad Yaghi had faced too and it led him to start DTEK.ai (previously Dukkantek).
The company aims to digitise the in-store shopping experience to elevate overall customer satisfaction. The flagship product SWIFT eliminates the possibility of waiting in a line at any retail location. The platform has been designed to be intuitive and fast and allows customers to check out up to 15 items in under 30 seconds.
The platform uses computer vision tech, where it intelligently identifies items and add them to the basket. For retailers this means increase margins by streamlining the checkout process and reducing overhead costs that are typically associated with the traditional checkout systems.
A report by Science Direct points out that the market size of global self-checkout systems are valued at $3865.8 million in 2022 and is expected to have a compounded annual growth rate (CAGR) of 13.4 per cent from 2023 to 2030. There globally there are companies like Crane Payment Innovations, Scandit, Leav and several others.
Nevertheless, there are complexities associated with self-checkout – machine errors, and even the loss of job for the cashier.
To combat some of these issues DTEK.ai already has worked on bringing in computer vision and AI technology. The first POS was aimed at digitising small grocery stores, this helped the team get a better understanding of the retail landscape.
“Dukkantek POS served as an invaluable learning platform, uncovering critical insights into the needs of the market and the operational hurdles retailers face. This ongoing dialogue between our team and the retail community as well as the customer behaviour insights, has enabled us to refine and enhance our core offerings,” explained Yaghi.
He added the team had realised there was need for flexibility and scalability in retail technology. This helps in a faster checkout experience, customers can expect to see a 70 per cent reduction in waiting times, and retailers benefit from 60 per cent higher traffic throughput.
“Responding to these learnings, we reimagined our core architecture to be more adaptable to diverse retail settings. This iterative growth process has been key to staying ahead, ensuring that SWIFT remains at the forefront of checkout innovation,” said Yaghi.
However, developing AI-based solutions present a unique set of challenges. Unlike established technologies, there isn’t a ready reckoner or playbook to follow. “The technology is so new that you have to blend creative problem-solving and risk-taking with a high speed of iteration to ultimately achieve the predetermined goal,” added Yaghi.
This meant the product and engineering teams had to devise innovative methods to construct robust databases that could support real-world applications. It was also important to keep the product economically viable. So what helped was working on first principles.
The value proposition was to increase speed without sacrificing accuracy. The first principle was to make sure the product was also priced in a way that made sense to the retailers, given their operational costs. This meant there was a need to focus on the hardware problems with good software.
“We couldn’t rely on expensive cameras and weigh scales to accomplish our goal. It would have made the product far too expensive to make any sense. As a by-product of that, if end customers didn’t like using it, as economically efficient as it may be, it wouldn’t drive the engagement that would ultimately translate into economic value for our retailers. So, it had to be fast, and it had to recognise items perfectly. Even 97 per cent wasn’t good enough, and three seconds wasn’t good enough. We are proud to say that our product now makes no mistakes and can detect as many items as you can fit onto it in under 300 milliseconds,” said Yaghi.
Currently the product uses AI and computer vision, a customer places multiple items on the SWIFT platform, and in under a second, each product is added – all the customer needs to do from there is tap for payment and head on their way out.
“Over 18 months, we carefully curated a team of experts and meticulously collected data to train and perfect our AI solution, ensuring that our technological solution exceeds expectations by being intuitive to use and as fast as humanly possible,” added Yaghi.
He explained the team is aspiring to be pioneers in transforming the retail landscape on a global scale and redefine how people interact with the retail environment around them. Specifically, the MENA region is undergoing rapid evolution, embracing modernization, and leading the charge in AI integration on a pace that is second to none globally.
Yaghi added the team wants to ensure that they position themselves to embrace and drive that change, actively contributing and shaping the region’s technological advancement.
“We aspire to be the pioneers in transforming the retail landscape on a global scale, redefining how people interact with the retail environment around them. We want to build best-in-class technology from the region, setting a new global benchmark globally, and offering solutions that are significantly more advanced than current standards,” concluded Yaghi.
