Professor Wallhäußer Award - Winner
The First Place of the Professor Wallhäußer Award for Innovations in GMP and Pharmaceutical Technology 2026 went to AstraZeneca Microbiology for their project "Automated Reading of Agar Plates for Environmental Monitoring with AI". Read the details below:
Automated Reading of Agar Plates for Environmental Monitoring with AI
In microbiology laboratories across the pharmaceutical industry, manually reading agar plates has long been a routine yet labour-intensive task. At AstraZeneca, this decades-old process has now been fundamentally reimagined. By leveraging artificial intelligence (AI), the company has successfully automated environmental monitoring plate reading—marking a significant step toward Pharma 4.0.
From Manual Work to Intelligent Automation
Traditionally, microbiology labs have spent thousands of hours each year visually inspecting agar plates and manually recording results. This approach not only lacks data integrity and the potential to retain permanent image records but also introduces subjectivity, as results may vary between trained analysts.
In collaboration with Clever Culture Systems, AstraZeneca adapted the “APAS Independence” system—originally developed for clinical microbiology—to meet the stringent requirements of pharmaceutical environmental monitoring. The result is the first scalable, commercially available solution tailored to high-throughput pharma laboratories.
High Performance and Accuracy
The system is capable of processing more than 200 plates per hour while accurately detecting microbial colonies. Crucially, it can distinguish real colonies from non-biological artefacts—overcoming a long-standing limitation of earlier imaging technologies.
It also automates data transfer directly into AstraZeneca’s global quality control platform, eliminating manual data entry and ensuring standardized reporting across sites. The system’s modular design allows it to adapt to various plate formats and monitoring requirements.
Validated to Meet Regulatory Standards
The AI model was trained and validated using more than 20,000 agar plates, incorporating a wide range of growth patterns, media types, and operational conditions. Validation followed international guidelines such as USP <1223> and Ph. Eur. 5.1, covering parameters including accuracy, precision, specificity, and detection limits.
Importantly, the system achieves the sensitivity required for aseptic manufacturing, where detecting a single colony-forming unit (CFU) per plate is critical.
Efficiency Gains and Cost Benefit
The introduction of AI-driven plate reading delivers substantial efficiency improvements. At large manufacturing sites, manual review is reduced to just 10% of plates; 1–3% of plates showing actual microbial growth and some false positives due to media artefacts and batch to batch variation.
At AstraZeneca’s Macclesfield site, where around 30,000 plates are processed monthly, the system is expected to save approximately 4,000 working hours annually. Across the global network, projected savings exceed $20 million over five years, with a typical return on investment achieved in less than two years.
Improved Quality, Traceability, and Compliance
Beyond productivity, the system significantly enhances quality assurance. Plate images can be digitally archived, enabling full traceability and retrospective analysis. Automated data capture reduces transcription errors and accelerates reporting times.
Compliance with regulatory requirements, including 21 CFR Part 11, is supported through comprehensive audit trails and secure electronic records.
A Blueprint for the Industry
With successful implementation across eight manufacturing sites, AstraZeneca has established a new benchmark in microbiological quality control. The system provides a scalable model for other pharmaceutical companies seeking to modernize their laboratories.
By combining advanced imaging, machine learning, and automated data integration, this innovation represents a turning point—demonstrating how AI can transform even the most established laboratory processes.