In high-volume order operations, pallet building drives a large share of logistics costs. How many pallets are needed, what goes on each one, and in what order they are loaded are decisions often made at the dock, on the fly. AI Cubic Planner, Janis Commerce's AI agent, moves that decision to the start of the process: it plans every pallet before picking begins. To learn about the other capabilities launched alongside the agent, read Product updates: from the first pallet to the end customer.
Why pallet building impacts logistics costs
Every pallet that ships with empty space is paid for as if it were full. That empty space translates into more trips to move the same volume, a higher cost per unit shipped, and dock hours spent rearranging loads.
On top of that, when pallet building happens at the end of the process, the result depends on the experience of whoever is building pallets on that shift. Someone with years at the dock knows how to figure it out on the fly, but the decision comes too late, when the order is already picked and fixing it is expensive.
What AI Cubic Planner is and how it works
AI Cubic Planner is an AI agent that plans pallet building before picking. Here is how it works in four steps:
- Analyzes each order: it takes the volume and weight of the products in the order.
- Defines the build: it determines how many pallets are needed, what goes on each one, and in what order they are built.
- Respects physical constraints: it accounts for maximum weight per layer, fragility, compatible categories, height, stacking, and the rotation options for each box.
- Guides picking: picking follows that plan, and each pallet ships exactly as planned.
Manual pallet building vs. AI-planned pallet building
| Built at the dock | Planned with AI Cubic Planner | |
|---|---|---|
| When the decision is made | At the end, once the order is already picked | Before picking |
| What it depends on | The experience of whoever builds pallets on that shift | A consistent rule applied to every order |
| Empty space shipped | Variable and hard to control | Lower, because volume is planned |
| Rework at the dock | Frequent | Lower, because the plan is ready to execute |
| Transportation planning | Defined once the load is already built | Planned with more certainty, because volume is known in advance |
What we learned while building the agent
Developing AI Cubic Planner was not a straight line. According to Leonardo Gambino, Head of Product at Janis Commerce, the first versions built pallets that were perfect on paper: they used every inch of volume and respected the maximum weight of each layer. In practice, however, warehouse operators rebuilt them anyway.
One variable was missing: the rotation options for each box, meaning whether it could be laid on one of its sides, placed upside down, or only rest on its base. Without that data, the model built combinations that worked in theory and fell apart at the first move. Adding it and recalibrating the logic is what made it possible to execute the agent's plan exactly as is, with no corrections.
Other new capabilities on the platform
AI Cubic Planner is part of a set of four capabilities recently launched in Janis Commerce, all built on the same idea: resolving in advance decisions that are made manually today.
Dynamic picking waves
Every time a Wave starts, pending orders are automatically reprioritized based on their committed delivery date. The most urgent ones go first and the rest move to the next wave, without losing the efficiency of multi-order picking. For a refresher on how this method works, read Types of Order Picking in Warehouses: Complete 2026 Guide.
Order cross-docking
When inventory is spread across stores, dark stores, and distribution centers, each order can use the network as a bridge to reach destinations with no stock. When it is received at a node, the next shipment is triggered automatically, with no manual intervention and end-to-end visibility.
Data protection
A permission defines which profiles can see customers' full personal data in orders and deliveries. Everyone else sees it masked. Information is visible only to those who need it, in line with regulations such as GDPR.
Frequently asked questions
What is AI Cubic Planner?
AI Cubic Planner is an artificial intelligence agent from Janis Commerce that plans pallet building before picking. It analyzes the volume and weight of each order and defines how many pallets to build, what goes on each one, and in what order, to reduce empty space shipped and rework at the dock.
What variables does it consider to build a pallet?
The agent considers the volume and weight of each product, maximum weight per layer, fragility, compatible categories, height, stacking, and the rotation options for each box. Combining all of these variables makes the plan executable in the warehouse with no corrections.
At what point in the operation does it step in?
AI Cubic Planner steps in before picking. Each pallet build is defined in advance and picking follows that plan. That way, the decision is no longer made at the dock, when the order is already picked and any correction costs time and money.
What type of operations is it designed for?
It is designed for high-volume order operations that ship on pallets, such as wholesale or B2B operations. In those cases, how each pallet is built has a direct impact on the number of trips and the cost per unit shipped.
How does it help reduce transportation costs?
By planning pallet building before picking, the agent reduces empty space shipped. That means fewer trips to move the same volume and a lower cost per unit shipped. And because the volume of each order is known in advance, transportation can be planned with more certainty.
Optimize pallet building with Janis Commerce
AI Cubic Planner joins the Janis Commerce platform so that good pallet building no longer depends on who is on shift and becomes an operational standard. If pallets in your warehouse are still defined at the dock, there is real room for improvement.
