Jacobi Robotics recently partnered with Delta Technology to help a defense manufacturer turn chaotic, injury-prone manual pallet building into a predictable automated operation. The production runs on a FANUC industrial robot.
The manufacturer operates a true make-to-order fulfillment workflow, building parcel pallets for freight carriers. It continuously accepts, packages, and releases customer orders downstream. Cases arrive to be palletized in completely random order, reflecting live demand, not prebuilt waves or optimized sequences.
This setup means the manufacturer doesn’t have the opportunity to pre-plan pallet composition or fix recipes that define what’s next. Pallets must be built on the fly, in real time, as cases arrive, while still meeting stability, safety, and shipping requirements, including boxes up to 50 lbs.
This kind of randomness is common in parcel shipping, and exactly where traditional palletizing automation struggles most. At the dock edge, this real-time variability translated directly into manual labor.
Operators were continuously lifting, twisting, and repositioning mixed cases to build outbound pallets. Over time, this led to an elevated injury risk from repetitive and awkward motion, increasing difficulty staffing palletizing roles, and a safety gap at the final step of an otherwise automated operation.
For a manufacturer in the defense industry, this was unacceptable. The decision was clear: Remove people from the most injury-prone task without sacrificing flexibility or throughput.
Jacobi provides a mixed-case palletizing system
Mixed-case palletizing is building stable outbound pallets from random, unsequenced cases—without predefined recipes or advanced knowledge of what comes next. To automate this step, the manufacturer partnered with Delta Technology to deploy a robotic mixed-case palletizing cell powered by Jacobi Robotics’ OmniPalletizer.
Two design choices are what made this viable in a live environment, Jacobi said. First, the system fits into the existing operation, with no sequencing required.
Instead of demanding upstream buffers, perfect flow, or a greenfield redesign, the cell drops into existing conveyor lanes and works with the case flow the building actually produces. This is the core “brownfield-native” premise: Upgrade the workflow, not the building, including supporting dozens of SKUs in daily operation (with no SKU limit in OmniPalletizer).
Second, the system included real-time planning for unsequenced case flow. OmniPalletizer continuously decides where the next case should go and how to place it safely when the stream is unpredictable. This is a “holy grail” problem the industry has discussed for years — building stable mixed pallets under real constraints, not controlled lab conditions.
“Mixed-case palletizing is one of the most injury-prone jobs in the warehouse, especially in outbound environments where variability is the norm,” said Max Cao, co-founder and CEO at Jacobi Robotics.
“Our goal is to remove people from repetitive, high-risk palletizing work without forcing customers or integrators to redesign their entire operation,” he said. “This deployment shows that you can automate mixed-case palletizing safely, flexibly, and in a way that actually fits real production.”
Inside the OmniPalletizer
This deployment highlights the full Jacobi mixed-case palletizing stack, from hardware to software to live operation. This stack starts with the mixed-case palletizing cell. This is a modular robotic cell built around a FANUC industrial robot, designed to integrate directly into existing outbound lines without requiring warehouse redesign, building two destination pallets configurable by load carrier.
Next is the operator interface (HMI). Jacobi said an intuitive HMI that enables operators to:
- Monitor pallet builds in real time
- Handle exceptions without engineering support
- Keep production running through shifts and demand spikes
- Operator training time is less than one day.
Once the robot is in action, users have access to a real-time view of it dynamically building mixed pallets from random case flow. Jacobi said the workcell can adapt continuously as orders are processed, achieving 100% stability on all pallets and up to 90% cube utilization—outperforming human stacking.
“The real breakthrough here is that the system doesn’t rely on predefined pallet recipes or advanced knowledge of the order,” noted Yahav Avigal, co-founder and chief technology officer of Jacobi Robotics. “We’re planning and validating every placement in real time, using physics-aware motion planning and continuous feedback, so the robot can make safe, stable decisions even when case flow is completely random. That’s what allows this cell to work reliably in live parcel palletizing operations instead of controlled lab conditions.”
The defense manufacturer reports initial gains
With Jacobi, Delta Technology, the integrator, was able to validate performance upfront using simulation rather than trial-and-error, with 0% error between simulated and actual cycle time. Additionally, the robot can avoid brittle, rule-heavy palletizing logic and behave predictably under real order fulfillment variability.
Results from the deployment include:
- Operators removed from repetitive, high-risk palletizing tasks
- Improved safety at the dock edge
- Successful automation of fully random, real-time parcel palletizing, handling thousands of cases per day
- Faster deployment with lower integration risk, including a 2-week commissioning duration and a 1:1 relationship between simulated and actual cycle time
- A scalable solution ready for future expansion, with dozens of SKUs in daily operation — and no limit on the number of SKUs supported — boxes up to 60 lb. (27.2 kg), and two destination pallets configurable by load carrier
- 100% stability on all pallets
- Up to 90% cube utilization—outperforming human stacking
- Operator training time is less than one day
With the system, Delta said it was able to treat mixed-case palletizing as a reusable subsystem, not a bespoke science project. This reduced commissioning time and project risk and included a two-week commissioning duration.
Save the date for the 2026 Robotics Summit
