Warehouse Plus Labs · Physical AI

Physical intelligence, built here.

Warehouse.Plus built this lab so intelligence has to survive the floor. It is not finished in a simulation. Its scope includes warehouse workflows, WMS and OMS context, and physical evaluation.

Where

Warehouse operating conditions

On what

WMS and OMS context

Then

Validation before scoped rollout.

Illustrative visualization

The lab

Physical intelligence is the work of this building.

  1. 01

    Warehouse Plus Labs is where the promise we make to the brands on this platform gets built. The promise is physical: pallets, paths, and orders, not a slide.

  2. 02

    The next warehouse infrastructure is built where software intelligence meets physical execution. That meeting is this lab.

  3. 03

    An evaluation needs explicit evidence and site context before a wider rollout is considered.

Illustrative visualization: The lab floor from above
Illustrative visualization

What it is

A Physical AI lab, built by Warehouse.Plus.

This is a physical intelligence lab. Its subjects include AI, robotics, and new operating designs in the context of warehouse work and existing systems.

Each idea needs a defined hypothesis, measurements, and an assessment of interfaces and operating conditions. Any rollout is scoped to the site.

What physical intelligence has to prove.

Three research directions, with evaluation criteria to define before a trial.

Illustrative visualization: A driverless forklift moving a pallet
Illustrative visualization

01 — Robotics & automation

Machines that earn the aisle.

From mobile robots to goods-to-person to agentic machines that decide on live inference.

Evaluation topics include travel, handoff, receiving, picking, and putaway in a multi-client warehouse context. A demonstration alone does not establish an operational result.

Illustrative visualization: A pick aisle ready for the next move
Illustrative visualization

02 — AI & operational intelligence

Decisions taken from the live floor.

Forecasting, slotting, replenishment, and pick paths, plus agents that act.

The questions include what systems should observe, recommend, and explain. Data access, decision boundaries, and evaluation conditions need to be agreed before a trial.

Illustrative visualization: Robot arms placing cartons on a conveyor
Illustrative visualization

03 — Digital twin modelling

The floor, and the model of the floor.

The arms meet the cartons. The model tries the same move before another site does.

Digital twins can model workflows, slotting, and facility design. A useful evaluation defines how the model relates to warehouse conditions and which evidence would support a next step.

Everything the lab is responsible for.

Physical intelligence is not one feature. These are the subjects a trial in this lab can be asked to move.

Autonomous mobile robots

Travel, handoff, and idle time on a live aisle, with other work still in the building.

Goods-to-person

Whether the person waits less, and whether the station still clears the order.

Agentic robotics

A machine that chooses the next action from live inference, then is scored on the action it took.

Driverless forklifts

Pallet moves on the real receiving, pick, and putaway the warehouse already runs.

Demand forecasting

The next peak, checked against what the floor actually shipped.

Dynamic slotting

Where the carton sits, changed while orders are still being picked.

Predictive replenishment

Stock moved forward before the pick face goes empty.

Pick path optimization

A shorter path that still finishes the order the customer placed.

Live WMS and OMS

The same systems production uses. The lab does not keep a private copy of the truth.

Real and synthetic orders

An evaluation may use agreed operational data or synthetic scenarios, depending on access and trial conditions.

Receiving through returns

Inbound, pick, pack, ship, and the return. A trial that skips the return is unfinished.

Facility design

Layout and flow stressed in the model before they are cut into another building.

How it works

A finding leaves the lab in three steps.

01

Hypothesis

Every experiment starts with a hypothesis, the success line, and the measurement. We do not run a trial without knowing what success looks like.

02

Real conditions

The work is receiving, picking, packing, shipping, and returns. Real and synthetic orders go through the live WMS and OMS. A vendor demo does not tell you how a machine behaves on your floor. Orders do.

03

Network deployment

Consider rollout only after evaluating the evidence, interfaces, and conditions at the intended site.

Illustrative visualization: Warehouse floor concept
Illustrative visualization

For the brands on the platform

An evaluation should inform what happens next.

The evaluation subjects include pick paths, demand forecasts, and inventory placement. Evidence from a trial should be assessed before any operational change is considered.

Faster paths mean faster delivery. Better forecasts mean fewer stockouts. More precise inventory means fewer errors at the moment that matters most.

Physical intelligence is changing the warehouse.

What separates the companies that lead from the ones that fall behind is the ability to build and scale real operational intelligence. Not a vendor demo. Not a hope that the floor will behave like the slide.

Warehouse Plus Labs is where we build ours.