Sometime in 2017, Paccurate made the leap from bin packing algorithm to artificial intelligence. No, it didn’t become self aware and start plotting to overthrow its human masters (we’re pretty sure). Rather, it gained the ability to make real-time decisions based on a data set far too complex for a human to consider.
Paccurate is a class of software called Narrow AI. Don’t let the “narrow” misnomer fool you- all that means is that we’re dealing with an AI that’s completely focused on a specific problem and not, say, Skynet. For example, IBM Watson is a Narrow AI, albeit a fairly complex one managed by a team of very smart people.
This new, smarter form of Paccurate didn’t suddenly emerge overnight. Instead, we began to gradually find more and more instances where its packing decisions seemed inexplicable or wrong. This was honestly disconcerting at first. Did we have a robot rebellion on our hands, or worse, a bug? Only after time-consuming analysis on a case-by-case basis were we able to pick apart Paccurate’s reasoning for some of its strange decisions.
It was always right.
Here’s what had happened: we had given Paccurate more and more complex inputs to consider, and the means to analyze them, before it could return an output. The result was a surprising number of completely unintuitive but nevertheless optimal packing suggestions. We quickly relished the unintuitive, and evangelized it to our enterprise customers. The stats speak for themselves:
An average 6-20% reduction in shipping spend on multi-item shipments
An average 13% reduction in corrugated cardboard (around 1 square foot of cardboard per carton)
Around 20% reduction in cubic volume of shipments, which means fewer trucks on the road and less carbon in the air!
AI is a tremendous tool for controlling costs, reducing waste, and bolstering corporate sustainability initiatives. There’s simply not time for a human to sit down and account for every single factor that might inform the right packing decision, and weigh them against each other. Here’re some of what you, a puny human, are up against, and what Paccurate considers in the blink of an eye:
Carrier rate tables punish inefficiency: No matter how well you’ve negotiated your carrier rates or dim factor, there are always incentives to pack a certain way. A shipment might be cheaper in one box if it’s going Zone 2, or two boxes if it’s Zone 9. Where’s the breakpoint where two boxes is cheaper? When is it cheaper to purposely incur a dim fee rather than split the shipment into two boxes?
Corrugated and fill are not free: Overboxing is the biggest point of material waste in the supply chain. Can you measure an extra square foot of cardboard against the rate factors above?
Labor costs: Is a box optimal if the product configuration is too complex for a human to assemble? Is one of the items in the order on the other side of the warehouse? If so, when does it make sense to split the order up vs sending the worker on a hike?
Business and physics rules: Just how fragile is that lampshade or vase? What is allowed in the box with it? Upon which axis can it be spun for safe transit? In what order should the truck be loaded so that it can be easily unloaded on a delivery route?
That’s a lot of questions. The bad news is that those are only a subset of things that should be considered for cost-optimal packing, stacking, or loading. The good news is that we have the technology to do it right.
Editor's note: I, for one, welcome our robot overlords.
In the good old days, packing optimization consisted of post-it notes or crusty old laminated signs in the warehouse, directing workers to which boxes to use for which products. If those post-its didn’t foresee a certain item combination, the hope was those workers would be able to rely on tribal knowledge and make packing decisions for themselves. It was a simpler, happier time.
These days, the “just eyeball it” approach is at best outmoded, and at worst devastating for a business's bottom line. Leveraging AI in your packing, stacking, and loading processes can have dramatic benefits. We’d like to show you how.
Consider upgrading your fulfillment process with Paccurate, the smart, real-time packing decision maker.