eel to machine learning Until recently, determining whether an item needed an extra Amazon box was a painstaking manual task, product by product. Teams would review details like whether something already came in a sturdy carton, how fragile it was, and how it might fare on a conveyor belt and in a truck. That approach worked, but at the scale of “hundreds of thousands of items”, as Aidan D’Souza, who leads Amazon Australia’s Ships in Product Packaging program, put it to Inside Retail, the time “quickly adds up”. The turning point came when the Australian packaging team partnered with the local AWS group to build an AI tool designed and deployed here first. The system ingests a wide range of inputs – item names, product descriptions, bullet points from vendor listings and images from fulfilment centres – and uses machine learning to estimate key product attributes. Those attributes then inform whether an item can safely be shipped in its original packaging with no extra box at all. In practice, this has been a step change in speed. In a single month, the AI-enabled process enrolled more than 12,000 additional products as eligible to ship in their own packaging – a body of work that D’Souza said “would have taken 18 months” under the old manual system. Overall, more than 150,000 products in Australia are now eligible to ship in original product packaging, with that number “growing” as the tool scales. What customers will actually notice For customers, the most obvious change is what arrives at the front door. According to D’Souza, one in 10 packages leaving an Australian fulfilment centre today now ships with no added delivery packaging at all. Many more orders come in flexible paper bags and padded envelopes that can go straight into household recycling bins, instead of single-use plastic mailers. Mailata saw this shift up close on the packing line. He described walking the warehouse floor and being struck by “the amount of original packaging for specific items that were just going out on the conveyor belt, straight into a trolley, trolley to the truck.” “Why would you put something that was already in a box in another box?” he recalled thinking. “That doesn’t make any sense.” For him, the value shows up at home as well as in the warehouse. He pointed to “super simple, efficient, easy” parcels that are quick to break down and recycle, without “extra fluff” like Styrofoam, bubble wrap or loose plastic bags. “You don’t have to worry about where is this all going once the package arrives,” he said. “It’s going to be all super simple to recycle it and put away.” AI as a sustainability accelerant While the partnership with Mailata shines a light on Amazon’s packaging ambitions, D’Souza is clear that the real story is the acceleration AI enables. The local tool doesn’t just save time; it reshapes what’s possible. By turning messy, imperfect product data into usable signals at scale, the team can make better packaging calls across hundreds of thousands of SKUs, not just a curated few. That matters when you zoom out to Amazon’s global packaging footprint. Since 2015, the company’s packaging reduction initiatives have helped avoid more than four million metric tonnes of packaging worldwide. In Australia, the AI-powered Ships in Product Packaging program is still relatively young, but already, Amazon has tripled the number of local orders that ship with no added delivery packaging since 2021. Globally, around half of customer orders now ship with less packaging or none at all, while in Australia one in 10 orders from a fulfilment centre is already delivered in its original packaging. D’Souza sees this as a textbook example of Andy Jassy’s public push to use generative AI and machine learning to make customers’ lives “better and easier”. Condensing 18 months of manual work into one month is, in his view, “just scratching the surface” of what AI can do for packaging decisions. Next on his roadmap: using the same technology to enrol even more products, even faster. Reducing waste, not protection Of course, “no added packaging” only works if products still arrive intact. Both D’Souza and Mailata come back repeatedly to protection as non‑negotiable. “Some things need extra protection to make it to the end zone,” Mailata quipped in the media release. “My role is to make sure Amazon can protect your package… but using the least amount of delivery packaging needed.” Inside the fulfilment centres, that balance shows up in material choices as well as box decisions. When packaging is needed, Amazon says it prioritises lightweight options and has phased out single-use plastic delivery bags in Australia in favour of paper-based solutions and paper dunnage that can be placed in kerbside recycling. Water‑activated tape replaces plastic alternatives and is designed to go into the same recycling stream. Customers are already noticing – and not just as reviewers. D’Souza, who spent three of his five years at Amazon in customer service, said teams actively monitor reviews, contact centre feedback and social media, as well as their own everyday experiences as shoppers. He joked that, as someone who orders coconut water weekly, not having to break down a box every time has made his life “a lot easier”. A cultural shift in the making For Mailata, who grew up in a big family where his mum’s obsession with tidiness and efficient packing shaped his own habits, the packaging project is as much about mindset as machinery. “I just want to spread awareness,” he told Inside Retail, especially among young fans watching him tackle waste as seriously as he tackles defensive ends. His message: take pride in recycling, pack efficiently at home, and understand that “your efforts as a small part in this world go such a long way in a big community”. For Amazon, Australia’s AI‑driven packaging work is already drawing interest from other regions keen to replicate the model. If that happens, the tool built here could help reshape packaging decisions across the company’s global network, accelerating a shift from “big box, lots of plastic” to something far more precise. And if Mailata is right, the simplest proof will be on doorsteps across the country: fewer boxes, less waste and parcels that still show up right on time.