Applied Integration and its sister company AST was approached by a UK-based water sampling and testing facility to improve the consistency and efficiency of a repetitive post-incubation screening process.
After an incubation period, samples must be inspected to identify potential growth indicators—such as spore activity or mould development—before determining the next handling step. The existing workflow relied heavily on manual sorting and visual assessment, which was time-consuming, labour intensive, and prone to variation between operators. With growing sample volumes, the requirement was clear: reduce manual handling, improve repeatability, and free scientists from routine screening so they could focus on higher-value analytical work.
Design Approach and System Development
Our goal from the outset was to automate the screening and sorting workflow while keeping the process simple, reliable, and compatible with the client’s established laboratory procedures.
The design centred around a self-contained automated cell capable of managing multiple samples at once, with a storage section that allows batches to be loaded and left running unattended. The system was built to reduce touch-time, minimise bottlenecks, and maintain sample control throughout the process.
A coordinated handling mechanism moves samples through each stage of the workflow, positioning them precisely for inspection and ensuring consistent presentation to the imaging system. This approach supports throughput while maintaining careful handling and traceability.
Automated Imaging, AI Decision-Making, and Sorting
At the heart of the solution is an AI Engine using camera imaging and trained models to assess each sample after incubation.
For every sample, the system:
- Captures a high-quality image
- Compares the image against trained AI models
- Determines the most appropriate outcome based on the classification result
From there, the handling is fully automated:
- Samples that require further inspection are retained and routed to a defined hold area
- Samples that meet defined pass criteria are automatically removed to waste via a bin route
This removes the manual, costly labour of routine screening and sorting, while ensuring every decision is applied consistently using the same model-driven criteria.
Performance Improvements and Operational Benefits
By automating image capture, assessment, and physical routing, the system significantly reduces manual workload and improves consistency across the screening process.
Key benefits include:
- Reduced scientist time spent on repetitive sorting and visual checks
- Improved repeatability through model-based decision making
- Faster throughput through unattended batch operation
- Reduced handling and improved workflow efficiency within the lab
Most importantly, the solution allows scientists to focus on complex analytical tasks rather than routine classification and movement of incubated samples.
Outcome
The completed system demonstrates how targeted automation and AI-based inspection can modernise a traditionally manual laboratory workflow. By combining automated handling, machine vision, and an AI Engine for consistent classification, the facility has moved to a streamlined process where only the samples that need attention are retained—improving efficiency, reducing cost, and increasing day-to-day operational capacity.
