The industrial digital transition leading to automated, connected and integrated manufacturing and supply chains is nothing new, as the new paradigm centered on the pervasive use of the Industrial Internet of Things (IIoT) and other enabling technologies is active since the beginning of the new century. 

The new element introduced by the Covid-19 emergency is represented by the boost the transition received due to the lockdown, and the consequent need to move many activities to remote. Smart-working has become a “new normal” in many instances, and social distancing impacted also on how the different operations are run in the industrial facilities. This trend of innovation is expected to compress in the next 18 months what would be normally done in five years, according to an article by Anna-Katrina Shedletsky published on Forbes. 

Companies are called to plan great investments for the digital transition, focusing on the creation of data-driven teams and IT infrastructures while leaving the legacy ERP managemet systems. The Boston Consulting Group suggests from its website to dedicate around 10% of AI investments to algorithms, 20% to technology, and 70% to business process transformation.

The new approach is not limited to Western economies: emerging countries are also investing in smart industrial technologies to maintain and increase their competitive potential, as acknowledge for example by the policies based on Industry 4.0, 5G and IoT the Vietnamese government put in place (see here the Vietnam Investment Review). 

Cloud systems to manage data

Cloud data storage is an enabling technology fundamental to manage from remote the great mass of data generated within the smart manufacturing sites and transmitted to the central Manufacturing Execution Systems (MES) and Warehouse Management Systems (WMS) by the capillary network of sensors, barcodes, RID tags, smart-gates reader systems, Real Time Localisation Systems (RTLS), smart devices, etc. 

A supposed lower security of data stored in Cloud systems would have represented the main barrier to the full implementation of this technology, says Shedletsky. This is expected to fall very rapidly, allowing for the full integration of data in a single repository along the entire supply chain, from where to monitor all internal and external activities from remote. Under the new perspective, thanks to data analytics it will become easier to identify and solve in real-time any issue arising in the production lines, plan inventories and maintenance activities and track in detail shipment and delivery of raw materials, APIs, intermediates and final products. 

The adoption of integrated digital manufacturing technologies shall prove critical with respect to the ability of delivering robust regulatory evidence and authorisation dossiers in the highly regulated pharmaceutical sector. Data integrity is becoming a central requirement, that can be easily handled thanks to smart manufacturing modalities able to capture and safely store data along every step in the product’s life cycle. 

An example of the Cloud potentiality in the R&D contest is represented by Benchling, a Cloud centralised platform developed by MIT’s student Sajith Wickramasekara to automatically track interlink related data, such as genetic sequences, cell lines, reagents samples, results, and experimental conditions. Data can be then elaborated to produce reports and exported to files to be used for regulatory purposes just in few days. Named by Forbes among the Next-Billion Dollar Startups, Benchling has just closed a new round of investment during the pandemic, for a total funding of $114 million and an overall valuation of $850 million.

How to overcome issues with social distancing and procurement

The pharmaceutical manufacturing cycle is already greatly automated, thus the impact of social distancing is not expected to be very relevant. Different might be the case in the administrative departments, or in other industrial sector (i.e. electronics) where human working was up to now more convenient than automation. The request to increase distance between operators may lead to new investments in automation to become a valuable option, according to Shedletsky. Many people that experienced the benefits of home working during the pandemic may also prefer to continue with this flexible modalities, leading to less workforce going back to physical workplaces.

The implementation of AI-based smart manufacturing sites may help solving the critical issue of APIs’ procurement from low cost countries (i.e.China and India), that greatly impacted Western’s pharmaceutical productions during the Covid-19 epidemic. A complete rethinking of the supply chain and the availability of “reduntant” manufacturing sites is becoming something worth to reason on, suggests the Boston Consulting Group. The final goal is to increase the company’s value chain through 24h/7d integrated manufacturing over multiple locations, so to respond to the increased request of “personalised” products. This is also true in the pharmaceutical field, there the 4P medicine paradigm and the advancements in gene editing and other biotechnological techniques is generating new “tailored-made” therapies requiring industrial capacity for small batch manufacturing. 

According to the Boston Consulting Group, the ability to rapidly scale up AI-based manufacturing systems will represent a critical factor to face the many uncertainties resulting from the Covid-19 crisis, including disruptions in operations and supply chain and changes in consumer priorities. The extensive use of AI to analyse data shall support better forecasts’ predictions and real-time decision-making, allowing for a greater flexibility of resources and improved cost efficiency along the entire value chain. The future may see a greater number of smaller productive plants located closer to the final customers in the different geographic areas. 

The current situation

The European Union has greatly invested in the digital transition, as confirmed by its new Industrial Strategy, and it represents the leading geographic area (51%) in the implementation of AI in manufacturing operations, followed by Japan (30%) and the US (28%). 

Maintenance (29%) and Quality (27%) are the main areas of AI application in the industrial contest, according to a report from Capgemini. Production (20%) and Product engineering/R&D also see a good penetration of enabling technologies, while supply chain management appears to be far behind (8%). Deep-learning based on data from previous machinery’s failures coupled to real-time process monitoring through sensors is at the core of AI-based planning of maintenance activities. Computer vision systems are able to analyse the single items moving along the production chain, selecting the low quality ones to be purged from the batch. This results in an improved quality of the production, paralleled by robust documentation of data. Procurement processes can also be greatly improved thanks to orders’ predictive planning based on historical data.  

The transition to fully smart industrial plants needs careful planning and it should be afforded using a step-by-step approach, starting from a small pivotal experimentation and set up of critical parameters in a selected area of operations. This smoothly leads to complement legacy and new smart systems, resulting in a robust governance of the process and integration with the IIoT. This first prototype can be then progressively extended to include all the different areas of activities, first in a single location and subsequently to include multiple manufacturing sites connected to a central AI platform for data management.