Codes & Streusel at digitalHUB Aachen: Where Agent-Based AI Meets Industrial Practice
March 17, 2026 | digitalHUB Aachen
When AI experts and industry users come together over streusel rolls and concrete use cases, it creates exactly the kind of exchange that makes a difference—not the next white paper, but direct conversation. The Codes & Streusel format at digitalHUB Aachen thrives on exactly that. On March 17, 2026, aiXbrain was there again—this time with a workshop, demo, and lively discussions on digital sovereignty in the AI era.
Keynotes: Agents as Teams, Processes Instead of Tools
Two inspiring keynotes set the tone for the event.
Vlad Larichev built on a now widely cited thesis by Nvidia CEO Jensen Huang: In the future, AI agents will collaborate just like human employees—and should therefore be managed as teams to fully realize their productive potential. A perspective that shifts the focus: away from individual models, toward orchestrated systems.
Kira Frings highlighted the challenges that arise when transitioning from individual AI tools to supporting entire processes—especially when it comes to meaningfully integrating AI with existing IT systems such as ERP solutions and their data structures. Anyone who has ever tried to connect an LLM agent to an established ERP system knows: this is where much of the actual work lies.
Demo: The AI Popcorn Machine Is Back
Tangible rather than abstract—that’s the vision with which aiXbrain's Marketing Manager Dirk Neuß brought the AI-powered popcorn machine back into the spotlight. The demo combines agent-based AI, the IXON Cloud data platform, and IoT sensors from Daqore into a fully functional system—demonstrating how industrial applications can be built from these components. Sometimes you just need popcorn to spark the idea.
Workshop: AI Solutions and Typical AI Problems
The workshop led by AI Solutions Manager Holger Pigerl was designed to spark discussion—and it did. The participants’ diverse backgrounds ensured that different perspectives quickly clashed: from questions about the evolution of LLMs and AI agents to very concrete use cases.
Some of the key topics:
- Simple vs. difficult use cases – Where are the low-hanging fruits, and where are the real challenges?
- Specific application areas – When does an LLM make sense, when an AI agent, and when a classic ML model?
- Vendor lock-in and digital sovereignty – How can we avoid dependence on individual, especially non-European, model providers?
- Self-hosting – Which models are well-suited for self-operation and for what purposes? What infrastructure and expertise are required for this?
- Security in non-deterministic systems – How can we reliably manage LLMs and agents?
- Limits of agent-based AI – What safeguards are needed to prevent worst-case scenarios?
The workshop concluded with a brief overview of the strategic prerequisites for successful AI deployment—organization and corporate culture, technical architecture, compliance—and a roadmap for AI projects with aiXbrain.
The biggest takeaway: Stability amid rapid change
The technological landscape is shifting virtually every week due to new foundation models and agentic frameworks. At the same time, uncertainty is growing: Are we relying on the right model? How dependent are we becoming on a specific provider?
For aiXbrain, this underscores a principle we consistently adhere to: developing clearly defined use cases based on concrete needs. Where are the real pain points—and what must the solution actually deliver? A well-thought-out data strategy and modular, open software architectures are investments with lasting value—regardless of which model dominates the conversation in six months. Vendor independence is no coincidence, but a design principle.
Conclusion: Personal, concrete, valuable
Codes & Streusel is not a conference format with glossy slides—it is a place where real questions are asked and real answers are sought. The team at digitalHUB Aachen has once again put together a professional and diverse event that facilitates exactly this kind of exchange.
Personal networking in the AI scene isn’t just a nice-to-have. It’s the channel through which trust is built—and trust is the foundation for AI projects that truly work.
We look forward to the next one.
