“At the pandemic’s peak, most of the time-series models in production failed to see the sudden surge/drop in demand.”
Sandip Bhattacharjee, Head of AI/ML at Tabsquare.ai
What’s one thing that is common between logistics and machine learning algorithms—optimisation. Cost optimization, inventory management and many other aspects of a delivery pipeline generate a stream of complex data. AI offers tools that can tackle this kind of data. While the pandemic has disrupted the supply chain, the logistics industry is bracing for a different type of disruption—Artificial Intelligence. For instance, e-commerce has seen tremendous growth with retail sales reaching a worldwide total of $4.28 trillion in 2020. The ease of buying everything online has given rise to delivery services across the nation to come up with innovative options for fast and efficient delivery. While the concept of online delivery is not new, the functioning of postal services has gone through a considerable transformation over time. It is no longer a matter of shipping things from point A to point B. India especially, is slowly becoming home to one of the fastest growing online markets in Asia Pacific.
“Retail e-commerce sales worldwide totalled $4.28 trillion in 2020, with e-commerce revenues expected to reach $5.4 trillion in 2022.”
Dunzo Digital, a Bengaluru-based hyperlocal delivery venture is using AI and machine learning to change the e-commerce playfield by building a postal service for Indian cities. The on-demand delivery service platform is using AI to figure out personalized content and options to decode the best approach to accomplish a task, and lastly, deciding on the best partner for order fulfilment. It partnered with Britannia and Cloudnine Group of Hospitals to deliver food, essentials and medicines to doorsteps during COVID-19.
Mailing system, too, has undergone significant changes thanks to machine learning. A large number of postal operators today rely on barcodes. Information about the senders and recipients are recorded through AI. In combination, the increasing prevalence of quick-sortation and automated sorters help in postal centers is improving the sorting speed and cost effectiveness, as well as the reliability and personnel requirements of sorting. Deutsche Post DHL Group, leading mail and logistics company, finds Indian an important market base. With the majority of freight movement on road, DHL has launched SmarTrucking in India to accelerate the development of technology-enabled logistics solutions. DHL uses Internet of Things (IoT) and data-driven insights to increase the return on time and resources. This provides almost 95% of total end-to-to-to-end shipment visibility, but 50% reduction in transit time compared to the traditional trucking industry.
Another supply chain service company,’Delhivery’ has grown to become India’s most prominent third-party delivery service provider. DHL offers a wide range of services in transportation such as express parcel, LTL and FTL freight, cross-border, B2B & B2C warehousing, and reverse logistics. The Delhivery AI platform ingests and assimilates 1 billion system actions, 30 million addresses, 4 billion GPS locations, and 2.7 million terabytes of computer vision data.To meet the demand of the ever-changing e-commerce ecosystem, Delhivery uses AI, ML and big data to create a flexible, responsive and adaptable delivery system.For example, Without addressing standards, users are typically inconsistent, which may include alternative spellings, as well as geographic information. Delhivery’s Addfix leverages machine learning techniques to disambiguate unstructured addresses into an addressable form. Another Delhivery tool ‘ ‘Catfight’ derives subcategory and category information from an unstructured product description.
In the matter of structured address, Amazon India is not far behind. Amazon is already spending $2 billion in Indian operations and has a long-term strategy using AI technologies. It has been using AI and machine learning to detect junk addresses, compute contact scores, and provide corrected addresses. There is no doubt that companies have become bullish about supply chain innovations. Whoever was on the edge were pushed over the line by the pandemic. Anomalies like COVID-19 help build better models in the future. But, the current deep learning assistance to the traditional time-series modelling is still underwhelming. “At the pandemic’s peak, most of the time-series models in production failed to see the sudden surge/drop in demand. On one hand, some product/service categories were seeing >10X demand due to stockpiling by end customers leading to a complete chaos in the supply chain,” said Sandip Bhattacharjee, Head of AI/ML at Tabsquare.ai when asked about tectonic shifts in the supply chain.
