Andrei Danescu: Dexory's Vision for Tomorrow's Supply Chains

Andrei Danescu (AD): The ultimate goal for Dexory is to transform the logistics industry operations and how it uses robots and AI-powered digital twin technology. We are working to become the standard data and insights solution for the supply chain industry globally. We want to provide our customers with cutting-edge technologies that help them to optimise their warehousing and supply chain processes, reduce their cost overheads and enhance their overall management through real-time, accurate data.

We have seen great demand for Dexory’s value proposition and we are addressing real challenges that supply chain operators have in terms of gaining real-time visibility of their warehouses and driving actionable insights for the industry.

(AD): For the autonomous robot itself, we have opted for a telescoping, extendable design that can go up to 12 metres in height, with cameras mounted all the way to the top, allowing us to measure the entire rack size in one scan. We also use cameras, lights and LIDAR sensors mounted at specific points of the robot. We use over 17 cameras on each robot for scanning and for manoeuvring the robot autonomously in the warehouse aisles.

As a full stack supplier, we can integrate with all the required systems that bring in relevant data points to enrich the insights and analytics we offer our customers. Systems include warehouse management systems (WMS), labour management as well as other types of management systems.

(AD): Achieving high levels of accuracy involves training on diverse datasets, real-world learning across a huge variation of warehouses and continuous improvements as crucial components. When our robots scan a warehouse, they produce incredibly rich data which is analysed by the DexoryView platform. The more scans that are carried out, the richer the data and the more accurate information the solutions can provide to businesses.

(AD): Simple answer is: Absolutely! How the industry has changed means that there is a need for real-time insights into the stock levels. This is particularly important as businesses need to understand their warehouse occupancy. Space is at a premium in the warehouse industry and for every £1bn that is spent online requires an additional 775,000 sq ft of warehouse space. Businesses also need to forecast their capacity better and receive suggestions for space maximization.

(AD): Dexory not only uses machine learning and computer vision to understand what the robot scans but also for autonomy to map, localise and avoid obstacles (static and, most importantly, dynamic). However, AI is key to analysing and providing forecasting and simulation capabilities. This technology can suggest efficiency improvements, new warehouse layouts, new ways of filling the space or storing goods based on seasonality, or the ideal number of customers on a site.

(AD): In general, any company that is developing AI solutions for any industry needs to ensure data collection quality, consistency, speed and how to integrate the data from various sources in real-time to drive insights and actions. In the logistics industry in particular, there is a growing need to achieve accurate data collection daily for the entire warehouse and to understand the depth of that data from 3D points, pictures. There are various diverse and large datasets that need to be gathered, understood and fed into models that adapt to real-world variability, such as delays caused by disruption, whether it is geopolitical or caused by unexpected disruptions.

Any AI-powered solutions for the logistics industry need to be able to evolve with the changing market demands and technological advancements. AI solutions need to be adaptive and allow for continuous improvements to stay relevant.

(AD): Warehouses remain the ‘black hole’ of supply chains as harnessing data is labour-intensive, inaccurate and outdated. Over 6000h/y are wasted on basic stock checks with 35% of the budget tied to the inventory costs of poor process optimization. 70% of available data is not being used as it’s inaccurate or siloed, meaning operators waste over 24 minutes to resolve a stock discrepancy. A typical warehouse operates at only 65% capacity with more than 20% tied up in obsolete stock, leading to millions in lost revenue. Any business with a warehouse will face these challenges - manufacturers’ procurement units, 3PLs, cargo handlers, and contract logistics/fulfilment companies.

These are just some of the challenges we help companies across the globe solve. This will allow warehouse staff to focus on key operations, which is important as the industry is grappling with a lack of resources when demand is at its peak. Using AI to access both real-time and historical data, provides companies with information to forecast and plan more accurately, improve their efficiencies and have smarter management of their workforce. We have recently helped organizations such as ID Logistics, Iron Mountain, Maersk and Menzies Aviation gain more visibility and actionable insights of their warehousing operations.

(AD): Yes we do. We can’t provide further information at the moment, but what I can tell you is that an important part of the roadmap is revolving around the latest AI technology to extract the last bit of insights from the data.

(AD): The industry needs to embrace the transformation that AI can deliver. The logistics industry has traditionally relied on historical data, but historical data does not provide insights. The impact of AI in warehouse management will drive efficiency and productivity, by automating tasks, reducing errors, enhancing productivity, optimising transportation, cutting costs and providing better visibility. As mentioned, AI is something that needs to be adaptive and allow for continuous improvement to stay relevant.

AI models need to be robust enough to handle the various scenarios and anomalies that are part of day-to-day operations in the logistics industry. From here, it is a matter of getting the right data, at the right time. This is key for ensuring that the technology succeeds in the warehousing space and delivers excellent ROI. Businesses need a simple way to contextualize the real-time data with information from various systems, through an intuitive and easy-to-use platform.