What if you could take a real operational problem from the rail freight industry and turn it into a working AI prototype in an afternoon? 

That’s exactly what happened at railXchange 2026 in Frankfurt. Next to the usual panels and presentations, Rail-Flow and Menlo79 ran a live AI hackathon under the theme rAIl Innovation Now. Customers and industry professionals didn’t come to watch a demo, they came to build one. 

How it worked

Melanie Maszke, AI Transformation Manager at Rail-Flow together with Nils Hoffmann, Senior Manager Growth & Solutions Engineering at Menlo79, were leading the AI Hackathon. She’d asked Claude to explain what a hackathon is to someone who’s never heard the term: “Explain to my grandma what an AI hackathon is.” The answer: think of it like a cooking competition — except instead of food, teams build digital solutions. Same time pressure, same need for creativity, same moment at the end where you find out whether it worked. 

The ingredients were simple: a real challenge, a team, and two tools Rail-Flow uses in its own daily work: Cursor for development and Figma for design mockups. Each team was guided by an AI team captain from Rail-Flow or Menlo79. No prior AI or coding experience required. The only prerequisites were openness and a willingness to try. 

Eight teams tackled four challenges pulled straight from the operational reality of rail freight: two teams per challenge, working independently, making it possible to compare not just the solutions but the thinking behind them: 

• AI-powered infrastructure document management: categorising emails about route changes and disruptions, and visualising them on an interactive map 

• Intelligent inbox automation: classifying incoming operational emails, routing tasks, and generating automated replies for standard message types 

• A marketplace for wagon storage: connecting track owners with operators needing temporary siding space, think “Airbnb for sidings” 

• A shift exchange portal for rail staff: replacing ad-hoc message chains with a structured, automated system for drivers, dispatchers, and yard teams 

The Results Spoke for Themselves

By the end of 150 minutes, all eight teams had clickable prototypes to show during the pitch. 

Some examples of those were the infrastructure monitoring app, that drew an immediate reaction from an attendee who wanted it integrated into her company’s existing system on the spot. The email automation prototypes showed that intelligent triage of rail operations inboxes is not a distant possibility, it can be set up in an afternoon. The wagon storage marketplace gave shape to a commercial model the industry has been circling for a while. And the shift exchange platform replaced a genuinely painful coordination problem with something clean and automatable. 

The Results 

The infrastructure monitoring app drew an immediate reaction from an attendee who wanted it integrated into her company’s existing system on the spot. The email automation prototypes demonstrated that intelligent triage of rail operations inboxes is not a distant possibility — it can be set up in an afternoon. The wagon storage marketplace gave commercial shape to a problem the industry has long acknowledged but never systematised. And the shift exchange platform replaced a genuinely painful coordination problem with something clean and automatable. 

The jury, which included Dr. Michael Beck (Founder & CEO, neXt Capacity), Burkhard Bräkling (Transport & Digital Consulting), Fabian Stöffler (Menlo79), and Dominik Fürste (Rail-Flow), and Babette had seen the full day’s worth of discussion about AI’s potential. What they saw in the hackathon was different: not potential, but evidence. 

One jury member put it plainly: “We have demonstrated that we can get really good results in a very short amount of time. And it’s repeatable.” 

What It Actually Proved

The format was designed to test a hypothesis: can rail freight domain experts, not AI specialists, produce functional prototypes in a constrained time window, using tools available today? The answer was clearly yes.  

Many of the participants came in with limited prior AI experience. The morning’s keynote by Christoph Zöller had already reframed the question: AI won’t take your job, but it will change what the job looks like, and the organisations that start now are building an advantage that compounds. The hackathon was the afternoon version of that argument: not a claim, but a demonstration. 

What made it work was not the tools. Cursor and Figma are powerful, but tools alone don’t produce results. What made it work was the combination of people who understood the operational problems deeply, the dispatchers, the planners, the logistics managers who live these challenges daily, and a structured environment that gave them permission to build rather than just discuss. 

That combination is rarer than it sounds. Rail freight has no shortage of conferences where operational problems are named and agreed upon. It has a significant shortage of settings where those same people are handed tools and two hours to try solving them. 

The Bigger Point

Speed matters, but it’s not the whole story. What the hackathon demonstrated is that the gap between identifying a problem and having a working solution in front of you has collapsed, when you have the right tools and the right people around you. 

Rail freight has always had complex operational challenges from inbox overload to fragmented infrastructure data and manual shift coordination. None of that is new. What is new is the speed at which those challenges can become working solutions. Not a concept document or a roadmap, but a clickable, testable prototype built in an afternoon, ready to put in front of the people who actually do the work. That’s the opportunity Rail-Flow is building toward. 

Interested in exploring what AI could look like for your rail operations?  Get in touch with Rail-Flow’s Transformation AI manager Melanie Maszke (m.maszke@rail-flow.com)