We sat down with inVia Robotics co-founder and CEO Lior Elazary to learn more about the last 12 months at inVia Robotics.
The Robot Family Grows
inVia Robotics had a notable year in 2019, with the deployment of over 400 robots to new clients and the growth of their surge (peak) fleet to support spikes in order fulfillment periods. inVia deploys autonomous mobile robots using robots-as-a-service (RaaS) and they also invested this past year in improving the management of their fleet of robots in the field. Their robotics operations center (ROC) is now solidly on a 24/7 support cycle and based in the US.
Now a Complete Warehouse Automation Solution
The inVia solution is now a complete system. inVia started with order picking tasks initially but in 2019 they added replenishment, sortation, put-back and cycle counting functions for the robots. This improves the ROI use case for the system and is opening the door to many new opportunities.
The side effect of extending robot capabilities is that it is enabling customers to keep a large fleet of robots working onsite during the year. Then, during surge periods, robots are taken off of tasks like replenishment, cycle counting or put-back and retasked for critical pick operations. This fleet role changing strategy helps to reduce the typical surge throughput to the range of an additional 20%, rather than 400%.
Last year, inVia also introduced the concept of a “put wall”, so that the robots can setup the tote order to optimize order picking by humans at the pack out station. This concept allows inVia to isolate processes in the interface between robots and humans. This is concept is evolving as inVia learns more about how pick order processing can be optimized within a warehouse.
Introduction of PickMate
In the areas of a warehouse where automation isn’t deployed, inVia introduced their PickMate software. Within PickMate, the inVia Logic software calculates the most efficient way to accurately take products from inventory to pack out for human workers. It operates via a tablet or smartphone, and communicates directly to the human picker workforce to enable them to pick, scan and deliver.
PickMate Screenshot – Image Courtesy of inVia Robotics.
The inVia Logic software uses the same artificial intelligence process to map and optimize a mixed workforce warehouse, and it allows a customer to deploy inVia Command to guide their people or inVia robots, or a combination of both.
The Launch of inVia Connect
The most notable product release for inVia in 2019 was the launch of inVia Connect, a new software solution that enables inVia to quickly connect their fleet management solution to a customers warehouse management system (WMS).
Lior said, “In the past, people approached it [integration] one of two ways: either APIs or batch data exchange. Once you have APIs, you can connect to our system, give us the information and we’ll do the work. And that’s how we started. The problem with that is that most people who own the warehouse don’t know how to work with it [the software] so they have to hire an integrator to write the translations between systems.” Lior went on, “The WMS has it’s own APIs, excel files exchanges, communication over sockets, or XML, or even an FTP server.”
He explained further, “We have a system that’s able to translate whatever APIs [or data] they have. So they take their APIs, they program that into our system. For example, JDA software communicates over iDocs. They drop it [a file] off on an FTP server, we pick it up. In the past you would create a separate module for JDA, or Scale, or Innerbank. What we’re encountering is that every singe customer has different needs, or different ways that they share information. Nothing is the same.”
inVia Connect helps with the mapping of messages, priorities etc, then it becomes possible to quickly get the information flow and business rules setup without the need for a system integrator. The result is that the integration happens in real-time, and inVia can now reduce the integration time to minutes and hours from days or weeks.
Application of Machine Learning
inVia is now using machine learning to understand the business rules/cases for order handling. For example, machine learning can understand if the orders fit a specific profile, and then change priority of an order or predict the ETA of a delivery based on real-time productivity factors within the warehouse. Pick order and robot scheduling in real-time is critical function for the inVia fleet manager. The result is that the system can effectively react to order batching patterns and determine the optimal operation of the robot pickers to deliver the right products to pack out stations.
The Changing Role of Integrators
In inVia’s viewpoint, integrators are evolving to be more involved in the configuration of the systems and in the implementation of the workflow than in writing code for the middleware between WMS and the robot fleet manager. inVia is finding themselves working more with consultants than with larger integrators.
At the Mobile Robot Guide, we also see this trend as a side effect of the success of the RaaS business model. RaaS innovators like inVia are delivering a complete solution to the end customer. In this case, that complete solution includes everything necessary to make the system function. With the introduction of inVia Connect along with AI-based workflow tools, there is no need for a classic integrator to be involved in developing or delivering the system. See our article on the Predictions for the Next Decade of Mobile Robotics for a deeper dive into this topic.
