- Event
- Prewave Day 2026, customer day
- Theme
- Resilience = Returns
- Format
- Keynote, 30 min, plus Q&A
- Audience
- Procurement, supply chain and risk leaders
We have never faced this many disruptions at the same time. Hormuz effectively closed, carriers back on the Cape route, tariff regimes rewritten by courts, rare-earth licences halted, the Rhine at Kaub down to 6 cm. Very few of these are black swans. Most are grey rhinos: visible, charging, and still ignored until they arrive.
The container changed global supply chains and made global trade possible. My argument in Vienna was simple: AI will have a similar impact on how we manage and operate supply chains, and on how we turn risk into resilience.
Five messages from the stage
- The new normal is a portfolio of grey rhinos. Geopolitics, trade policy, physical chokepoints, cyber, climate and insurance are moving at once. Risk management built for the occasional exception does not scale to permanent exception.
- Two forces shape the network of 2040. Geopolitical fragmentation and automation maturity span four scenarios. The primary case is Fragmented & Autonomous: regionally differentiated supply chains that use high automation to absorb the cost of network duplication. The downside is Stagnant & Exposed: fragmentation without the technology, run on manual firefighting and high buffers.
- The economics have flipped. For the same 1,000 EUR you bought about 50 million tokens with GPT-3 at the end of 2022. Today it is around 50 billion. Analysing three years of demand history for 10,000 SKUs takes roughly 15 million tokens, about 3 EUR. Cost is no longer the argument against.
- From firefighting to foresight. Today most of a risk manager's time goes into manual data assembly, reporting and ad-hoc war rooms. AI takes over that work and hands the time back for early warning, scenarios and diversification. GenAI also opens the black box of planning systems: a planner asks "Why do we have excess inventory in Europe?" and gets the cause and a recommendation in plain language.
- Onboard AI as a digital colleague, not a magic tool. A new colleague needs onboarding, a mentor (a "Meister"), first tasks and feedback. AI needs the same, plus context. An experienced planner who was burned once by the factory's unannounced August holidays pads the lead time every summer. A statistical model needs more than 100 delayed deliveries to learn the same. Context is the value.
To make the point tangible, I showed the end-to-end planning app I vibe coded in 30 hours. The same 30 hours could build a risk cockpit on your own supplier data. Do not wait twelve months for IT.
The slides
All 16 slides. Use the arrows or swipe, click a slide to open it in full size, or download the complete deck as PDF.
