Everyone’s talking about tariffs again. New hikes on Chinese EVs. Retaliatory measures on semiconductors. A growing push to reshore and de-risk supply chains. The headlines say supply chains are getting smarter. But on the ground, most global companies are still operating with limited visibility.
“We’ve made forecasting advanced,” says Johannes Herter, part of the founding team at Rome AI, “but many organizations still struggle to understand what’s happening in real time. Without that, even the most sophisticated predictions are unreliable.”
Johannes helped lead the development of intelligent systems that address this problem at the root – by rebuilding the fragile data foundations that global logistics depend on. He focuses on the foundational layer- often invisible, yet critical – that ensures the data reflects reality.
The trade world is shifting fast. The systems running it aren’t.
Johannes didn’t start in logistics – his background spans mathematics, machine learning, and system design, with degrees from ETH Zurich and early work at Mercedes-Benz R&D on intelligent mobility systems. After completing his studies, he joined Harvard University’s Visual Computing Group, where he worked on generative models that embed structure and meaning into dynamic 4D scenes- building on prior research in physical simulation that was later adopted for films like Elemental. That same systems thinking shaped Rome’s approach from the beginning applying it to the operational complexity of global supply chains. He’s rebuilding the systems that haven’t kept up.
Making sense of the world’s messiest data
“Supply chains operate like distributed systems,” Johannes notes. “But instead of clean APIs, they run on PDFs, emails, and mismatched spreadsheets.”
That becomes a real problem when the world starts shifting, as it is now. A tariff takes effect, and no one knows which shipments it applies to. A supplier changes freight lanes, and two teams record different delivery dates. A plant delays a shipment, and no one spots it until the customer calls. At Rome, he designed the foundational layer that reconciles disparate, often conflicting data from suppliers, carriers, and internal systems. It links purchase orders to shipments, flags contradictions, and constructs a reliable, unified view of current operations.
“It’s the kind of system that only gets noticed when it fails,” says Michael Hartmann, who designed Rome’s AI Infrastructure. “Johannes ensures it doesn’t. He made the foundation solid, so Fortune 200 companies can secure their supply chains without second-guessing the data.” This foundational infrastructure supports the rest of Rome’s AI stack: adaptive forecasting, exception detection, and scenario modeling that can operate under rapidly changing conditions.
A new foundation for supply chain AI
When asked what excites him about AI in logistics, Johannes emphasizes the importance of real-time clarity over abstract predictions.
“The true value of AI is in generalization, taking what it has learned and applying it to novel situations,” he says. “But that only works if the system first understands what’s happening now.”
Many global firms are still running critical operations on outdated files and fragmented tools. Rome addresses this by creating a consistent, machine-readable view of operations, one that systems and people can rely on to make informed, accountable decisions.
Though often behind the scenes, his work powers the core infrastructure that enables intelligent supply chains to operate with resilience and transparency. And in a world where trade policy, climate events, and geopolitical shifts can disrupt operations overnight, that kind of clarity may be one of the most essential tools global businesses can have.
Leave a Comment