Caja Robotics offers smart machines for the warehouses of tomorrow

From the Spanish word for ‘box’ or ‘present,’ Caja offers a flexible set of solutions in an age when COVID-19 is making online shopping bigger than ever.

A Caja picker station, to which the robots bring boxes to (photo credit: ASAF RAVIVO)
A Caja picker station, to which the robots bring boxes to
(photo credit: ASAF RAVIVO)
“People really hate doing warehouse work,” Caja Robotics CEO Ilan Cohen told The Jerusalem Post on Sunday. “The warehouses are huge and, due to the time pressure, workers have to run back and forth across huge distances and climb ladders and pick up heavy boxes.” Human error, such as dropping packages or delivering the wrong parcel to the dispatch center, costs warehouses time and money. 
Companies often rent or buy existing warehouses instead of building them from scratch, meaning that if the floors are not level or can’t bear a lot of weight, these are the conditions the company must work with and maximize.
Currently, between 90%-95% of warehouse labor is done by humans, and the hard work means that managers often overhire, as they predict workers will quit.
At the moment, Cohen says, the US alone needs a million and a half pickers, the people who remove items from one container and place them in the outbound shipment.
COVID-19 made a difficult job even harder, as workers might infect one another, damaging productivity even further. 
What Caja offers is a flexible solution involving two robot models, a smart algorithm, and a dash of blue-and-white originality from Binyamina. 
A robot can’t know what’s in a box, so it needs the Warehouse Managing System (WMS) to “tell” it what’s there. If the WMS “says” there are 40 shoes, the robot is unable to lift the box and think “this is too light for 40 shoes.” This is why humans, not robots, make excellent pickers. A human is able to see that a shoe is the wrong color or that something is broken; a machine cannot.
Robots must bring boxes as quickly – and safely – to the human pickers as possible. This is done with a cart robot and a lift robot. The lift robot can extend to deliver packages stored on a high shelf and the cart robot can deliver them.
As robots don’t get tired and AI can process a great deal of information in seconds, the system can predict what boxes will be in demand next. For example, maybe it’s Black Friday, or maybe the system “knows” a peak in orders is coming following Superbowl commercials. The robots will place boxes closer to the pickers’ station to save time when the orders pour in.
Some people treat robots as fairy dust, meant to solve all problems. Yet even the smartest robots won’t be able to drive fast if the floor has holes in it – nor should they, as they might hit human workers. Nor can robots “stop” performing a task they are told to do. In that sense, robots don’t make mistakes, programmers do. 
Caja robotics offers a flexible solution as their system allows a warehouse to start small, and see if performance improves or not. Clients also get access to a simulation that predicts what would happen when solutions are employed. The New Jersey based Bergen Logistics reported a 50% increase in productivity after they installed the system, Cohen says. 
The AI also sends the robots to a human worker that replaces their batteries once they reach 25% energy levels. “We’re working on having a robot do that as well,” Cohen explains.
Unlike other firms, vice-CEO Hanna Yanovsky told the Post, Caja Robotics develops the AI here and also builds the robots here, making the company blue and white at a time when the government seeks to invest in Israeli-made products.
Caja means ‘box’ or ‘present’ is Spanish, but in Hebrew, it also stands for “that’s the way to do it.”