The artificial intelligence added to parcel machines can already predict with 96% accuracy which packages will be picked up from the parcel machine within 24 hours. With a learning predictive model, packages can be delivered without unnecessary driving and to a location close to the consumer.
The filling of vans is limited by both the number and size of individual compartments: when small packages can be placed in cabinets of different sizes, larger packages can only fit in some of the cabinets. Hundreds of vending machines are completely filled every day or at least several days a week.
“Popular parcel vending machines cannot be expanded if they are located in a confined space, such as inside a store. With AI and machine learning, we can get a more accurate forecast of the capacity of parcel machines, which helps us to route packets more often to the parcel machines chosen by the recipients. In addition, more packages can be made to fit in certain package trays, ”he says Jari Paasikivi, Project Manager for Postal Parcel Routing.
The packet routing system uses artificial intelligence to anticipate the development of the available capacity of the parcel machine and predicts how many packets from the parcel machine will be retrieved before the driver brings in the next parcel load. The filling rate of parcel machines was previously based on a fixed average, which did not take into account the variation of the parcel machine, day or season.
The predictions made by artificial intelligence are almost 100% accurate
More accurate capacity forecasts speed up the process and allow packets to be delivered to recipients as quickly as possible. Artificial intelligence can already predict with 96% accuracy which packets will be picked up from the parcel machine during the first 24 hours.
“The package can be delivered as close to the customer as possible, which reduces unnecessary driving and the resulting emissions. We already know with almost 100% accuracy which of the individual cabinets in the parcel machine will be emptied and we will be able to book the right number of shipments for the next day, ”says Paasikivi.
We are already working on the next development project related to the use of artificial intelligence in packet routing. Before the package arrives at Posti, the online store will send a pre-notification. The time it takes to send a message varies depending on the online store. Artificial intelligence is used to predict when a packet actually arrives for sorting. If the time of arrival can be accurately predicted, advance reservations can be made for vending machines and the wishes of customers can be better taken into account.
“We have built our ability to use data for a wide range of purposes on a long-term basis. Artificial intelligence as part of our parcel machine network is an excellent example of how data and technology can be used to improve the customer experience. We have invested especially in our machine learning expertise, and the artificial intelligence that routes packets is entirely developed by our own personnel, ”says Riku Tapper, who is responsible for Posti’s data and automation.
Source: Suomen Posti