Follow the AI Leaders

Companies of all sizes are scrambling to unlock the benefits of artificial intelligence for their own operations. But it’s primarily the largest enterprises—those with the resources to pursue innovation—that are leading the way, using AI to retool their supply chains and make them smarter, faster, safer, more resilient and greener. In the process, they’re providing a glimpse of what the future holds.

FROM UPSTREAM SUPPLY chain functions like procurement to getting products into customers’ hands faster downstream, leading manufacturers, retailers and logistics providers are leveraging AI‑powered solutions to reap gains in productivity, speed and accuracy in fulfilling orders, sustainability and worker safety.

Just a few years after ChatGPT made generative artificial intelligence a household term, these companies are going big on investments in applications of both generative AI (GenAI)‑based technology and agentic AI. The technology is already bringing new levels of autonomy and efficiency to factories, warehouses and distribution centers.

GenAI is now viewed as the top technology powering end‑to‑end supply chain orchestration—the model of seamless integration across the value chain from sourcing of raw materials to point of sale, according to a study released last summer by procurement software developer and consultant GEP and North Carolina State University’s Supply Chain Resource Cooperative.

“The organizations leading the AI race are embedding GenAI into the core of their operating models,” the study stated, citing demand planning, warehouse management, category management, supplier contract management and sourcing as the top five supply chain functions where GenAI is expected to have the greatest impact.

While many large enterprises have been developing their own proprietary AI capabilities in‑house for years, they also serve as the test bed for emerging AI‑enabled technologies currently being piloted and refined by technology solutions providers and on the verge of becoming more widely available. A few examples:

• At its warehouses nationwide, PepsiCo is reaping significant productivity gains using a new AI‑based “decision agent” from MHI member AutoScheduler. The tool helps improve and automate decision‑making, thus reducing labor costs and standardizing performance across its distribution network.
• Walmart has turned to MHI member Symbotic to reimagine, develop and implement hundreds of robotic in‑store order‑pickup centers in the back of retail locations nationwide, leveraging AI‑enabled automation to expedite fulfillment of digitally placed orders.
• MHI member GreyOrange, whose AI‑driven warehouse orchestration system is already being deployed by top retailers and third‑party logistics (3PL) companies like Kenco and GXO, is piloting a new hardware‑agnostic AI solution that GreyOrange says will slash autonomous mobile robot (AMR) deployment times.
• Global industrial leaders like Smurfit Westrock already are enhancing safety with MHI member Blaxtair’s AI‑driven pedestrian detection system. Soon, a new feature will let managers query a GenAI agent in natural language to easily manage plant or warehouse safety with data‑driven insights.

In addition to deploying AI in their own supply chain planning, many of the largest enterprises are developing AI solutions to help their partners better manage their own supply chains. Walmart and Amazon, for example, provide tools to suppliers that forecast demand and manage inventory more accurately, thus helping the retail and e‑commerce giants minimize excess inventory and avoid stockouts.

AI solutions providers like Blaxtair CEO Franck Gayraud say the AI applications now in use are just the tip of the iceberg. The company, for example, expects to work with major forklift manufacturers to pilot a new AI‑based forklift trained to operate autonomously.

“It will be a revolution,” Gayraud said. “In our business, it’s going to change the way our customers operate their warehouses and vehicles in many different ways, and it will happen little by little.”

PEPSICO’S BIG AI GAINS

When food and beverage giant PepsiCo undertook a productivity planning exercise a few years ago, the company’s senior manager of warehouse orchestration, Peter Hall, decided it was time to do something different and look at the challenges dock leaders face in keeping operations on an even keel.

The challenge, as Hall described it in a webinar earlier this year, seemed straightforward enough: “How can we get the right case on the right trailer at the right time at the lowest cost?” But making the right decisions all the time isn’t so easy, given the chaotic and constantly changing nature of warehouses and varying experience levels of different supervisors at a single location and across all facilities.

The solution, developed by AutoScheduler, a provider of an AI‑based warehouse decision agent that harmonizes the multiple data streams that impact warehouses or plants—whether from the warehouse management system (WMS), enterprise resource planning (ERP) system or transportation management system (TMS)—and then optimally sequences tasks while supplementing critical decisions made by supervisors with unlimited query ability and automating more routine decisions.

“We harmonized all of the data across their operations and created dynamic orchestration plans for all activities, which resulted in a 9% productivity increase and provided 36‑48 hours of future-state visibility for all of leadership to easily see what is going to happen at every single operation across their network uniformly,” Keith Moore, founder and CEO of AutoScheduler, said.

PepsiCo’s productivity gain, which Hall said reflects the increase in the number of pallets moved per labor hour, is ” just unheard of in the warehousing space,” according to Moore, who calls the agent’s domain‑specific query capability “basically ChatGPT for warehouses.”

“Keep in mind we’re not actually changing the flows of the warehouse, we’re not doing anything besides making decisions more efficiently, so it’s either work harder or work smarter,” Moore said. “We are the work‑smarter piece of that equation and working smarter, there’s nearly 10% in an operation effectively that you can gain in throughput just by making the right decisions at the right time continuously.”

WALMART TURNS TO NEW AI‑BASED E‑COMMERCE SOLUTION

As the largest retailer on the planet with 2024 revenue of $675.6 billion and nearly 11,000 stores worldwide, Walmart has long been at the forefront of cutting‑edge technology to support its operations, turning to companies like Symbotic for AI‑based solutions to improve demand forecasting and inventory management, automate supplier management and squeeze costs and greenhouse gas emissions out of its transportation system.

Symbotic, which has worked with the retailer since 2017 to bring AI‑enabled robotic automation to its warehouses, is now designing in‑store accelerated pickup and delivery (APD) centers that will automate and speed up fulfillment of online customer orders. The agreement announced earlier this year calls for Walmart to deploy the new Symbotic systems in 400 APDs initially, with the option to add more in the future.

“Symbotic’s APD solutions will enhance the speed and accuracy of online order fulfillment, while offering new, convenient pickup options directly in the store for the consumer,” said Brian Alexander, Senior VP of Commercial at Symbotic. “APDs will reduce the logistical burden of small‑order fulfillment on distribution centers and minimize the in‑store labor needed to pick items for these orders, freeing up the retail workforce to focus on more complex revenue‑generating tasks.”

In effect, Walmart’s APDs bring the AI‑enabled automation solutions that the retailer has used in its warehouses for years directly to their consumers, adapting the technology for the microfulfillment space.

LEVERAGING THE LATEST ADVANCEMENTS

Typically, the companies leading the way in AI adoption for the supply chain also tend to be first in line to test new applications from technology solutions providers that incorporate GenAI or agentic AI capabilities—and the first to benefit as products emerge from the pilot phase and become market‑ready.

Expect these leaders to continue to weave AI into the fabric of their supply chains as solutions providers roll out new iterations of their products based on ever more powerful large language models (LLMs), the AI engines trained on vast amounts of domain‑specific data to understand queries in natural language, generate text answers and—in the case of agentic AI—execute tasks without human supervision.

Blaxtair, whose pedestrian detection system currently provides users with a dashboard where they can view safety data and key performance indicators (KPIs) collected by the AI‑assisted camera vision technology mounted on forklifts and other vehicles, expects to offer a GenAI‑based application by 2026 that will allow managers to ask questions of the system in natural language to uncover potential safety issues that need to be addressed.

Additional advances expected in 2026 will expand the system’s ability beyond pedestrian detection to report on any anomalies in the warehouse—a pallet blocking a door, for example—that could pose a safety hazard, Gayraud said.

Similarly, GreyOrange, the maker of warehouse optimization software, is looking to bring to market in early 2026 GreyMatter DeepNav, an AI‑powered solution for dynamically managing and optimizing autonomous robotic operations at scale, representing a “fairly big jump” in the company’s multi‑agent orchestration capabilities, according to GreyOrange CEO Akash Gupta.

Developed in collaboration with Google Cloud, the new product is intended to streamline the time‑consuming and expensive task of adding or adjusting the use of automated mobile robots in a company’s fleet. GreyOrange’s current estimates suggest that the technology will cut AMR deployment times up to 80%, even when managing large fleets of AMRs from multiple vendors.

AutoScheduler’s new “decision agent” is a standalone addition to the company’s core decision optimization platform and became available to the broader market in early October after being proven through use by PepsiCo and other select “power users.”

While the core platform previously could initiate a workflow into a WMS only with manual approval, the decision agent provides the capability to automate the workflow without human intervention.

“It is a true agent,” Moore said. “You used to have to go into a user interface and right‑click on ‘Approve’ and then it would go and trigger the workflow and do all of the work and release it on the floor. Now an agent can automatically do that.”

Symbotic, which, in addition to using AI and machine learning to continuously improve the routing and accuracy of its intelligent SymBots and fully integrated solutions, is now leveraging AI for demand planning and forecasting and also exploring how it can use AI to compress the time‑to‑value for clients when deploying warehouse automation systems.

“As a leading full warehouse automation integrator, original equipment manufacturer (OEM) of robotic technology and AI company, we see firsthand the extreme positive effects AI has on many companies’ bottom lines. There are a lot of robots out there that have fantastic physical capabilities, but if the intelligence of the robot—and the full system it lives in, for that matter—is limited, it will directly limit your gains,” Alexander said. “Just like running a team, if you hire the best and have them work together, collaborate and learn, they will start to predict, plan ahead and deliver more effectively with higher quality. Robots with static intelligence will always fall behind.”

Click here to read the full article.