AI Hype vs. Reality: Everyone claims AI will transform industries, but in logistics, where margins are tight and complexity is brutal: how much of today’s AI (hype) is actually usable in logistics? More reality than the sceptics admit – but less than the vendors promise. In logistics, AI works concretely today in three areas: communication automation, route optimization, and document processing. Everything else – autonomous end-to-end decisions, self-learning supply chains – is still in development, not product yet. We invest where the use case is already live, not where the pitch deck vision ends.

The Real Bet Behind HappyRobot: WaVe-X invested in HappyRobot to automate communication, coordination, even negotiation areas traditionally owned by humans. Are we underestimating how fast entire job categories could disappear? Yes, we’re underestimating the speed and abilities. HappyRobot doesn’t just automate communication, it replaces an entire coordination layer that today is built by humans with headsets, Excel, and a lot of patience. The question isn’t whether entire job categories will disappear, but how quickly companies will realize they’re already filling those roles worse than an AI could. The huge opportunity that most don’t see is that you can’t just improve what is there, but also enhance your teams with agentic colleagues. If your team is doing X calls, does that mean there aren’t more calls, or is it just that they can’t manage more? Would more calls perhaps lead to more volume? One can try these hypotheses today pretty easy and find out and that is the beauty of it in my eyes.

Trusting Machines with Money & Decisions: Would you trust an AI agent today to negotiate prices or manage payments on your behalf? If not, what still fundamentally breaks? For defined scenarios – yes, already today. Discussing a price within a limitation works fine and the AI has the advantage that it can hold the same conversation with X  counterparts at the same time and interchange the information given inbetween calls and steer the discussion accordingly. For open negotiations with a real counterpart – not fully yet. What breaks is the accountability. When an AI agent strikes a bad deal, nobody wants to own the liability. Not legally, not internally. That’s not a technical problem, it’s a governance problem – and no better model alone will solve it. But that is what I tell most colleagues: What can be automated will be automated – what differentiates us as humans is the ability to build real relationships that is where we need to focus. If something goes wrong – humans need a human to be mad at.

Europe’s AI Problem: Europe is pushing hard on regulation with the EU AI Act, while the US is scaling AI companies at speed. From your perspective: is regulation becoming Europe’s competitive disadvantage in industrial AI, or could it actually turn into an advantage? Both, regulation could long-term become a quality signal – “European AI” as a trust mark in critical infrastructure and industrial applications. Short-term, however, compliance costs time and capital that European scale-ups simply don’t have. The real problem isn’t the regulation itself – it’s that we still don’t have a European hyperscaler capable of turning it into a moat.

Winners, Losers & Timing: Five years from now, who in the logistics and industrial space will look completely outdated because they underestimated AI? The losers are already decided – they just don’t know it yet. Freight forwarders still relying on manual dispatching. Brokers without a data advantage. Mid-size 3PLs without a tech budget and without the courage to transform their own business model. In five years, survivors will either have become a platform themselves or will be running white-label on someone else’s. There’s nothing in between in my point of view.