Australia Post to resume work on AI-powered parcel lodgement kiosk

Australia Post intends to resume its evaluation of a smart kiosk for post 'shops' that allows customers to lodge packages themselves without staff assistance.

Credit: meldCX/AOPEN

Concept SALi - Smart Automated Lodgement API - was commissioned by Australia Post and developed by Melbourne-based meldCX with a handful of other partners.

It was unveiled last year at the annual National Retail Federation (NRF) show in the United States.

The self-service kiosk is built on AOPEN commercial hardware, and uses AI technology from Intel and cloud services from Google and Microsoft. 

“Concept SALi kiosks use powerful machine learning and computer vision technology to scan and detect each package - automatically verifying its size, weight, handwritten information, the sender's identity and shipping cost,” meldCX said at the time.

“This removes the guesswork, manual entry and complexity from shipping a package for customers and post office employees.”

General manager of the post office network at Australia Post, David McNamara, was quoted as saying that “Concept SALi makes it much easier and faster for customers to lodge parcels and provides more features that customers want, such as bulk lodgement.”

iTnews understands that when Covid-19 hit, Australia Post put the business case for development and adoption of Concept SALi on hold.

However, it is understood that Post now intends to resume work on the project.

An award submission to the 2021 Paris Design Awards mounts a strengthened business case for the smart kiosks post-Covid.

“With increased retail online spending during the pandemic, Australia Post is working hard to continue to deliver its services with no delays,” the submission states.

“As cities gradually reopen post-pandemic, Australia Post is looking to reduce person-to-person contact and increase post office efficiency, by shifting the parcel delivery lodgement process to a digital-based model.”

The submission says the smart kiosk “removes the guesswork, manual entry and complexity from shipping a package for customers and post office employees.”

The kiosk is capable of recognising handwritten labels with 95 percent accuracy and scanning in a package for delivery in under two minutes.

In production, the kiosks could also reduce queue waiting times at AusPost shops, with the submission claiming a 67 percent reduction from 15 minutes down to five minutes in peak times.

A video released to LinkedIn at the end of last year by meldCX shows Concept SALi in action, with a particular focus on how machine learning models were trained to recognise the details of parcels.