31 - AI : We can see clearly now

During the COVID pandemic, when international supply chains had never felt more volatile, it surely made sense to err on the side of caution and start stockpiling medicines to prevent shortages.

The reliance on China for APIs had created a single point of failure, prompting questions about whether manufacturers should move from just-in-time (JIT) to just-in-case (JIC), or repatriate more elements of production.

But rushing to this solution, even during a pandemic, isn’t necessarily best. Drugs and raw ingredients – particularly biopharmaceuticals – have expiry dates and often need to be stored in temperature-controlled conditions.

Blanket stockpiling or repatriation would increase production and storage costs to the point where critical drugs became unaffordable for governments and healthcare providers. Our aspiration remains the same, what we need to do is ensure that we’ve assessed and included risk properly in our models.

Holding firm on JIT isn’t about taking a cavalier attitude though. In fact, I’d argue that the opposite is true. The promise of big data, AI and machine learning (ML) is that it allows manufacturers to both stick to lean principles and reduce risk to supply in an uncertain world. The phrase I always come back to is ‘see clearer, see further, think smarter and act faster’.

In some circumstances, like a pandemic or winter flu season, companies might still stockpile since the data suggests that demand is highly likely to grow. But that’s very different to simply stockpiling because they assume demand will rise during this period. They’re still sticking to their JIT ambition – but good data modelling and appropriate risk assessment provides the evidence they need to understand the real, not perceived, risk and make sound evidence-based decisions.

In an increasingly digital supply chain, with real-time visibility, companies can extend their horizon to see whether events like lockdowns in China are likely to impact the production schedules or capacity of their raw ingredient suppliers. Everyone in the supply chain can only move as quickly as those behind them – but seeing further allows them to ‘act faster’ and mitigate impacts.

With more data, and better ways of interpreting it, manufacturers are able to move from static to dynamic modelling, so they’re continually re-evaluating the risks, and adjusting their orders accordingly. They could use digital twins to create a replica of the entire supply chain and gain real-time intelligence that allows them to proactively forecast demand and identify bottlenecks.

Using AI and ML to analyse large amounts of data also brings broader benefits to pharma companies beyond inventory and risk management. For instance, it can support continuous manufacturing, allowing teams to optimise production processes and run at the finest margin.

COVID exposed vulnerabilities in our global supply chains, when an aspect of the supply seemed resilient and turned out to be fragile. However, manufacturers’ ability to dynamically model and understand the risk is becoming more sophisticated all the time, putting them in a stronger position to manage future crises.

You can never eliminate supply chain risk completely – but seeing clearer, seeing further, thinking smarter and acting faster can certainly help to find the constantly-shifting balance point in a dynamic world and ensure that just-in-time doesn’t become just-too-late.