Business Happen 2026: How AI Agents at IT Sonix Deliver Measurable Value
On September 17, 2026, it was that time again: Guests from a wide variety of industries gathered at the „Digital Excellence“ premises of our parent company, Akkodis, at the Alte Messe Leipzig for IT Sonix’s Business Happen. Our series of networking events combines intellectual insights with a culinary experience, inspiring connections, and a shared perspective on future-oriented topics. This year’s theme: „From AI Hype to Agent-First Reality.“.
Because as of this year, one thing is clear: AI is no longer just an assistant—it’s becoming an active player. Companies at the forefront are relying on agents that work independently, connect systems, and structure processes on their own. This is exactly where IT Sonix comes in—with the goal of turning experiments into reliable solutions.
AI specialist and thought leader Max Mundhenke kicked off the program with his presentation „AI Readiness Instead of Knee-Jerk Reactions—Why Successful Agent Systems Start with the Right Use Cases.“ His take: The technology is becoming agent-ready, but most organizations are not yet. About 80 % of companies report improved individual productivity, but only about 37 % report a positive impact on EBIT. So we’re scaling AI faster than we’re scaling its value.
The key question, therefore, was not „What kind of agent can we build?“ but „What tasks should be handled differently here in the future?“ Mundhenke presented three prerequisites for effective agent systems: Process Readiness, Data Readiness, and People Readiness. He made it clear: A good use case describes work, not technology. Those who perceive automation as a threat rarely provide the best process data. Participation is therefore not an optional extra, but a prerequisite for technical quality.
Dr. Danny Hucke, AI Lead at IT Sonix, then demonstrated what this looks like in practice in a presentation titled „Reliable AI Solutions Thanks to Measurable Quality.“ His starting point was an experience that many participants immediately recognized: Pilot projects often seem promising, but they are often unstable or unreliable. Productive AI solutions therefore require a robust data foundation and, above all, measurable quality.
He illustrated this using three projects from IT Sonix’s day-to-day operations. A classification solution in the public sector that automates a task previously performed manually using the dual-control principle, and whose accuracy was iteratively improved from approximately 93 to 99.5 %. An AI-powered document processing system in the healthcare sector that processes sensitive data locally, recognizes text from scans that are sometimes difficult to read, and links insights across document boundaries. And an AI agent for company classification, whose assignment accuracy was gradually improved from about 88 to 97.5 %—compared to a manual error rate of three % in the previous process. IT Sonix also relies on agents internally: Our Wiki chatbot accesses the company’s proprietary knowledge base via agent-based RAG and runs on a local language model.
The takeaways summed it up: 100% % accuracy can never be guaranteed with probabilistic components. Where traditional software is sufficient, it should be used—only then should ML be employed, followed by LLM. Furthermore, every solution requires evaluation against real ground-truth data. Artificial intelligence delivers real added value precisely when the security and quality standards of traditional software align with the measurable quality of AI components.
Afterward, we handed the stage over to our guests: In an open panel discussion, participants joined Max Mundhenke and Dr. Danny Hucke to discuss the topic, ask questions, and share perspectives from their own industries. No “Business Happen” without snacks—a delicious buffet and cold drinks accompanied the evening until around 9:30 p.m., offering ample opportunity to reflect on the ideas discussed and make new connections.
A heartfelt thank you to Max Mundhenke and Dr. Danny Hucke for their valuable insights, as well as to all the participants for this inspiring evening. We’re already looking forward to the next Business Happen.
