While software has evolved to manage individual tasks, fleets of robots, and inventory flows, artificial intelligence promises to orchestrate all of them plus workflows, said GreyOrange Inc. Now, warehouse automation platform providers are in a race to provide such a holistic approach, noted Saurabh Gupta, who became chief technology officer at GreyOrange last year.
“I don’t come from the supply chain world,” he acknowledged to Automated Warehouse. “I’ve worked with NVIDIA, Amazon, and Apple, but the common theme has been robotics and AI, from consumer and medical to self-driving tractors.”
“After 20 years, the need for automation has become so much more clear,” said Gupta. “Before, everyone was trying to come up with use cases for their technology. People know that they want more autonomy, but they still don’t know exactly what they want.”
Founded in 2012, GreyOrange said its GreyMatter, gNetwork, and gStore offerings provide real-time visibility into omnichannel nodes and orchestrate robotic agents, people, and systems. The Suwanee, Ga.-based company said its hardware-agnostic systems enable customers to reduce their cost per unit, eliminate lost inventory, ensure worker safety and productivity, and enhance in-store experiences.
GreyOrange pivots from hardware to software
Over the past four years, GreyOrange has been working on GreyMatter and moving from providing autonomous mobile robots (AMRs) to focusing on its software suite.
“The company’s founders had the foresight to realize that, even as one of the largest AMR suppliers, the hardware will be commoditized,” said Gupta. “Software was limiting what happens in the warehouse. They decided to open up the platform to third parties to guarantee performance — that’s what we’re good at.”
GreyMatter works with robots from multiple providers, and GreyOrange now has projects with Hai Robotics and more than 15 other companies, he added. It works with suppliers and integrators through the Certified Ranger Network (CRN).
“Any company you can think of that’s building robot arms, sortation systems, AMRs, or humanoid robots — we’re in different stages of getting them on our platform,” said Gupta.
Hardware-agnostic platform opens opportunities
If its software and AI can manage different types of robots across environments, is GreyOrange limited to warehouses?
“The warehouse market is huge, and we’ve only captured a small part of it so far,” Gupta replied. “There’s significant value to be added, such as how to handle live operations with foundation models. That’s when you start to unlock a lot of value.”
“That said, nothing keeps us confined to the warehouse,” he continued. “We’re now deployed in manufacturing facilities — not just shipping and packaging items but also retrieving components and presenting them for assembly. That’s one place where we’ve seen a lot of traction, but the majority of our installations will still be in warehouses.”
AI can accelerate benefits, says GreyOrange
“AI progress is accelerating, and customers are starting to see benefits,” Gupta noted. “There are two core areas where we’ve made progress.”
The first is overcoming AI siloes, where one agent optimizes navigation and mapping, another one manages inventory allocation, and yet another handles order processing and sequencing, he said.
“Replacing all of that with a single model is already proving to have great advantages,” explained Gupta. “For example, with past models, inventory was allocated, and where robots were sent led to congestion. The answer was to think through navigation, where suboptimal inventory could actually lead to more throughput.”
GreyOrange has enabled heuristics to bring that information and developed a single AI model with information across orders.
“That’s a big foundational shift,” Gupta asserted. “We’re one of the few companies in a position to orchestrate from end to end. One working on just palletizing won’t have that level of control.”
The second area of focus for GreyOrange is applying large language models (LLMs) to look at vast quantities of data to identify choke points, complementing more classical pretrained models. The company has run robots in 100 warehouses over 100 billion miles and decisions, bringing it to an inflection point, said Gupta.
“There’s a tension between general-purpose and custom automation,” he observed. “With our DeepNav initiative with Google Cloud, we looked at warehouse parameters and found that those models are good at some things but bad at others. We’re not replacing everything but are combining more predictable models with LLMs. The two trains are converging, and we’re able to look ahead to possible outcomes to decide what to do right now.”
Gupta said that other areas where AI agents are starting to be helpful is in making coders more productive and expanding the capacity of human warehouse managers. “There are still too many manual tasks today, and AI is bringing decision making to managers’ fingertips. I don’t see humans ever being completely replaced; they can provide that last layer of input for goals.”
GreyMatter simulator provides predictive management
At MODEX in Atlanta last week, GreyOrange launched the GreyMatter Foundry simulator, which it said is designed to meet the growing need for AI tools to manage complex, heterogeneous warehouse environments including fleets of robots from different vendors, other forms of automation, and human associates.
“Warehouse automation should not be a leap of faith,” Gupta stated. “By putting the intelligence of our live, global GreyMatter network behind every simulation, we give distributors, 3PLs, retailers, and integrators a crystal ball grounded in real-world data.”
“Whether you’re designing from a greenfield scenario or rethinking an existing operation, Foundry lets you test thousands of scenarios, stress-test for peak demand, and arrive at deployment day with confidence,” he added.
The company has also entered a strategic partnership with Dematic in which the KION company will offer GreyMatter to provide customers scalable systems for distribution and fulfillment.