Leveraging Artificial Intelligence for Enhanced Inventory Management of Single-Use Solutions

Timothy Korwan, Director of Technical Applications; Alex Joyner, Digital Lab Product Manager; and Sarah Ann Landry, Operations Manager, Fluid Handling Solutions, at Avantor.

The use of single-use (SU) technology for pharmaceutical and biopharmaceutical manufacturing is one of the most important and transformative trends in the industry. It is estimated that the global market for SU technologies in biopharmaceutical manufacturing will grow from $4.3 billion in 2021 to $7.3 billion by 2026 with a compound annual growth rate (CAGR) of 11.3% for the period of 2021-2026.

SU manufacturers are now challenged to be more agile and flexible in their design, customisation and production processes — including assuring that their supply chains are sufficiently secure and globally diversified to meet their end users’ demands. As true system integrators, they draw from a broad range of raw material suppliers to create unique customised fluid handling pathways and systems.

These raw materials go into a “library” of SU components and include multiple types of tubing, filters, connectors, and sensors. In addition, there are critical products such as clean room supplies and personal protective equipment (PPE), as well as associated packaging materials, that must be fully stocked and managed to ensure production and delivery commitments are sustained.

Single-use manufacturers need powerful inventory management, production forecasting and control tools that will support the flexibility and certainty required to be agile and responsive suppliers for biopharma customers.

Challenge: Major Single-Use Inventory and Supply Chain Issues

Building and sustaining the inventory and supply chain resources necessary for productive and agile SU manufacturing is forcing SU providers to solve several interrelated issues. The first is lack of complete, real-time visibility into inventory of both SU raw materials and production-related materials. There are many state-of-the-art inventory management platforms that can, if properly established and with proper inventory control, provide the visibility SU manufacturers need to have, especially for regulatory compliance.

Secondly, single-use manufacturers continue to be challenged with unpredictable supply chains. The impact of COVID on global supply chains and lingering inefficiencies for critical components have led many manufacturers to adopt less efficient and expensive “just in case” overordering and supply hoarding practices.

While adding some elasticity to the supply chain instead of relying on extremely tight just-in-time ordering and inventory management practices makes sense, there are ways to make much more effective use of advanced inventory management and forecasting tools that use machine learning and artificial intelligence (AI) capabilities.

Solution: Leverage Artificial Intelligence (AI) For Data-Driven Decision Making

There is a growing recognition that the use of AI and machine learning tools can begin the process of moving from predictive, human-driven inventory management and manufacturing processes to genuine, data-driven decision making that determines, prescribes and essentially decides how best to leverage the supply chain and nearly perfectly match it with demand.

With AI and machine learning, SU manufacturers have stronger tools to automate and proactively identify deficits or roadblocks in both onsite inventories and supply chain resources. This can be done using newly available inventory tracking tools like smart shelves that use Internet of Things (IoT) technology to gather critical, real-time data.

This cuts down the time and effort required by sending personnel to manually scan shelves or double-check storage areas to track down materials that are supposed to be in stock. With AI learning and tracking what’s available — and what isn’t — production planning and inventory restocking becomes prescriptive; order replenishment can even be automated.

Data-driven decisions need as much data, in real time, to make effective choices that manufacturers trust. There are a range of IoT systems to generate that data, not only to guide current decision making but also to build up deep historic data that AI technology can mine to reach the ultimate goals of automated decision making.

Smart Buttons

Located throughout the workspace, these simple devices can be used by personnel to request quality control (QC) or engineering support from outside the clean room, enabling significant time savings. Providing a single point of contact for onsite support teams further standardises the process, and minimising the frequency with which personnel must enter the controlled environment reduces contamination risks.

Smart buttons improve throughput during assembly processes while increasing assurance that all critical QC inspections are conducted before an assembly can move to the next step in the process. Over the long term, machine learning can track when and why these requests were made and correlate with other data flows to identify ways to further streamline processes, potentially even automating support requests based on the long-term data captured.

Smart Shelves

These tools are crucial to transforming SU inventory management. The real-time data they provide contributes to building the critical insight AI needs to learn and ultimately prescribe order and replenishment cycles. Smart shelves have been shown to dramatically improve inventory visibility — helping minimise inventory discrepancies, reducing stockouts and generating detailed consumption data.

When fully integrated with ERP systems, they can support automated reordering of more than just SU components: high-volume inventory items such as PPE products, as well as all the packaging materials needed for safe, aseptic shipping of complete SU assemblies, are tracked and managed through tools like smart shelves. They also generate significant labour savings by reducing the need to have personnel inspect and scan shelved inventory to update and verify stock availability.

AI-powered Vision Systems:

Implementing an array of vision systems, both in storage areas and in clean rooms where SU products are assembled, can significantly improve quality monitoring and control, especially related to the proper use of PPE materials, as well as critical QC monitoring of SU assembly steps. Combined with data from smart shelves, they enhance inventory management, tracking and replenishment processes.

Through AI-driven vision systems, quality and specification assurance can provide critical backups through proactive deviation alerts if there has been an error in these steps. They also can contribute to AI efforts to harness data to better understand when and how errors in assembly and gowning procedures occur and help fuel prescriptive end user, inventory and market forecasting.

As single-use technology in biopharma continues to evolve, AI-driven inventory management for the full range of products needed — including components, PPE and packaging materials — provides a powerful tool to enable SU manufacturers to meet the demand.