Professor Wallhäußer Award - Second Place
The Second Place of the Professor Wallhäußer Award for Innovations in GMP and Pharmaceutical Technology 2026 went to Bayer / Berlin-Chemie / Merck Healthcare for their project "MADDOX: AI-Based Assistance Software to Increase Production Efficiency in a GMP Regulated Environment". Read the details below:
MADDOX: AI-Based Assistance Software to Increase Production Efficiency in a GMP Regulated Environment
As pharmaceutical manufacturing grows more complex and tightly regulated, companies like Bayer AG are facing a mounting challenge: how to maintain efficient, high-quality production with fewer experienced workers.
Demographic change, labour shortages, and increasing staff turnover are eroding the pool of skilled employees who traditionally hold critical operational knowledge. At the same time, stricter regulatory requirements and advanced technologies—such as track-and-trace systems—are adding layers of complexity to production and packaging processes.
When disruptions occur on the factory floor, machines typically report only symptoms, not causes. Identifying and resolving underlying issues often depends on years of hands-on experience. Without that expertise readily available, companies face costly downtime, higher waste rates, and delays in maintenance.
Traditional knowledge management tools, such as databases and digital manuals, have struggled to fill this gap. In time-sensitive situations, operators rarely have the capacity to conduct manual searches, particularly when terminology varies or technical expertise is limited. As a result, workers continue to rely heavily on direct support from experienced colleagues—an increasingly scarce resource.
To address this challenge, Bayer has introduced an AI-driven assistance system called MADDOX, developed in collaboration with Fraunhofer IVV and technology partner Peerox GmbH. Instead of requiring manual input, the system continuously analyses live machine data via standardized interfaces and compares current conditions with historical cases.
Based on this analysis, MADDOX delivers context-specific recommendations—ranging from text and images to video guidance—directly to operators via mobile devices and workstations. The system acts as a first-level support tool, helping employees diagnose issues faster, reduce downtime, and improve onboarding for new staff.
Crucially, MADDOX does not automate decisions or intervene in production processes. Operators remain fully responsible for evaluating situations and implementing solutions. A structured governance framework ensures that all knowledge content is reviewed, approved, and clearly labelled, maintaining compliance with strict pharmaceutical standards.
Since its pilot launch in 2019 at Bayer’s Leverkusen packaging site, MADDOX has been scaled across more than 70 machines on 13 production lines. Integrated into existing systems such as SAP maintenance tools, it has become part of daily operations for production, engineering, and maintenance teams.
The results are tangible: fewer repeat disruptions, reduced downtime, and measurable relief for experienced specialists. Equally important has been strong user acceptance, driven by a design approach that emphasizes support rather than surveillance or performance evaluation.
The system’s impact extends beyond Bayer. MADDOX has also been implemented at Merck KGaA and Berlin-Chemie AG, where similar benefits have been observed. Today, it is in use at more than 20 companies, highlighting its scalability across the industry.
As pharmaceutical manufacturers seek to balance efficiency, compliance, and workforce challenges, MADDOX offers a glimpse of how artificial intelligence can be applied responsibly—supporting human expertise rather than replacing it, and ensuring that critical knowledge remains accessible when it is needed most.