MHI Solutions Improving Supply Chain Performance
Physical AI, Agentic AI, Digital Twins and the Dawn of Autonomous Logistics
AS MIKE GEYER, head of industrial digital twins at NVIDIA, prepares for the future of supply chain operations, he believes they’re set to become “living ecosystems.”
How so, you may ask? According to Geyer, each physical process will be “mirrored and optimized through digital twins.” Not to mention, they’ll also be “orchestrated by agentic artificial intelligence (AI)” and “carried out by intelligent machines in the real world.”
“Together, these technologies will transform supply chain operations from slow, heavy and opaque to efficient, flexible and autonomous,” he said.
Pras Velagapudi, chief technology officer at MHI member Agility Robotics, agrees. He adds that future supply chain operations will be sophisticated ecosystems of physical AI, agentic AI and digital twins for one primary reason: Each technology fills a different gap.
For instance, physical AI enables robots to perceive and move safely throughout supply chain operations. Meanwhile, agentic AI enables operators to coordinate decisions not only across fleets but also throughout their facilities. And, finally, digital twins offer a “real‑time simulation layer” that connects everything.
“They form an integrated system that’s far more adaptive and efficient than any single technology on its own,” Velagapudi explained.
While working at MHI member Siemens Digital Industries Software, Christian Schmidt, director of technology, Digital Logistics, has personally witnessed the recent evolution of supply chains—from “linear, reactive systems into interconnected, intelligent ecosystems.” As he’s worked firsthand with global manufacturers and logistics providers, he’s discovered the evolution was absolutely necessary. Why? According to Schmidt, “the complexity and volatility of modern logistics demand it.”
“The convergence of these three technologies addresses fundamental challenges that traditional approaches cannot solve,” he said. “Digital twins provide the foundation—they create virtual replicas of entire supply chain networks that mirror real‑world operations in real time.”
On the other hand, agentic AI builds on the foundation that digital twins create, adding autonomous decision‑making capabilities. Agentic AI doesn’t just analyze data or offer recommendations, after all. Instead, it can execute complex, multi‑step processes independently.
From there, physical AI will execute each decision in the real world. As the “embodiment of intelligence in robots and automated systems,” according to Schmidt, physical AI must work together with digital twins and agentic AI for key reasons.
“Supply chains span both digital and physical domains,” he said. “A digital twin can identify that rerouting a shipment will reduce costs by 15%, while agentic AI can autonomously negotiate with carriers and update delivery schedules.”
He added, “Physical AI can adjust warehouse receiving schedules and robotic picking sequences accordingly—all within minutes of a disruption.”
The benefits of the combination have been considerable so far. Schmidt has found that organizations that deploy digital twins alongside AI report up to 20% improvements in fulfilling their customer delivery promises. Additionally, their labor costs have declined by 10%.
“Our own customer implementations show freight cost reductions of 10% to 16% and warehouse efficiency gains of up to 40% when these technologies work in concert,” he emphasized.
LOOKING AHEAD
Through the integration of simulation, autonomous reasoning and robotics, Geyer thinks supply chains will provide faster turnarounds, more resilience and higher resource efficiency, each of which “simply wasn’t feasible before.”
Velagapudi agrees with Geyer’s sentiment. In his opinion, the combination of physical AI, agentic AI and digital twins will “make supply chains dramatically more resilient, efficient and predictable.”
“By merging real‑world robotics with intelligent decision‑making and simulations, operations will be able to adjust to disruptions instantly, optimize continuously and run with fewer bottlenecks and less downtime,” Velagapudi said. “The merger will also mitigate the job gap that’s currently holding the industry back.”
While preparing for the future, Schmidt notes that supply chain performance will be fundamentally transformed across numerous dimensions, including resilience, speed and cost efficiency.
When it comes to resilience, he has noticed a direct impact from his company’s software offering. In particular, the offering, known as Supply Chain Suite, will enable supply chains to become proactively resilient rather than react to supply chain disruptions after they’ve occurred. By leveraging physical AI, agentic AI and digital twins, the offering will help Siemens’ customers optimize their logistics processes, resulting in greater efficiency.
Additionally, this combination is expected to improve supply chain operations substantially. For example, through physical AI, warehouse robots can scan up to 10,000 pallet locations every hour. In turn, they’re capable of providing higher real‑time inventory visibility than ever before, as the need for manual stock takes has been eliminated.
“When integrated with AI‑driven demand forecasting and route optimization, these systems enable same‑day production and delivery—a capability that was infeasible with traditional approaches,” Schmidt said.
Regarding cost efficiency, Schmidt has also noticed a variety of benefits, again due to the combination of physical AI, agentic AI and digital twins.
“Our customer base reports freight cost reductions of 8% to 18% through network optimization and inventory reductions of 8% to 10% through improved demand forecasting and allocation,” he said. “It also reports warehouse operating cost reductions of 10% to 15% through AI‑powered space optimization and robotic automation.”