The bane of all supply chains is uncertainty in demand or supply.
Just-in-time supply chains work great when there is little disruption. Lately, disruptions seem to be the norm rather than the exception. The more tightly linked the supply chain, and the closer you get to just in time with minimal inventory, the faster any disruption propagates through the system.
Everyone got burned badly during the pandemic. The immediate visceral reaction was “If we only had inventory, we could have mitigated this.” Along with “We need more visibility” and a multitude of other things including scenario planning. As the pandemic subsided the focus has shifted to building resilience and robustness within the supply chain.
Companies use inventory to buffer against disruption/uncertainty and increase service levels because it allows the organization to decouple. But as any practitioner knows, when inventories increase, so do costs and lead times.
The debate about whether you can increase resilience by working with many suppliers to mitigate risk or fewer suppliers to streamline the supply chain boils down to tradeoffs relating to inventory, probability of failure, the cost associated with failure and the complexity of the supply chain.
There is tremendous pressure to reduce costs, while decreasing response times and increasing service levels in every organization; this is nothing new.
Within supply chains, the sheer volume of data (zettabytes and yottabytes), the number of possible scenarios, the degree of uncertainty and the magnitude of the impact is starting to rapidly increase. This, along with the need to get answers quickly, is pushing the limits of what a human being can reasonably be expected to compute. As Michael Mullany states, large-scale data and content analysis has come and gone in waves because of the “inexorable expansion of scope and size of the data that we want to analyze.”
Hence, companies are reaching to AI in an attempt to tackle this. It’s not an easy task, because knowledge is embedded in the minds of people. Broadly, some of the areas where AI is actively being looked at because of the physical human limitations are:
- Manual number crunching of data from different sources
- Reaction time to changing data
- What-if scenario crunching
- Generating actionable analytics
- Discerning patterns before they are obvious
- Creating end-to-end visibility
- ATP dates and volumes
For AI to become practically applicable, use cases must be identified, models developed, refined, tested, decisions have to be validated, all of which takes time. So, the adoption of AI into supply chains will cost more and take longer than what most companies are anticipating. It will still require human beings to be at the controls, because at least for the moment HAL-9000 isn’t quite here yet for supply chains.
The upside is that there are things possible today that weren’t in the realm of possibility in the not-so-distant past. Brilliant minds and companies are pouring money and effort into creating solutions. Which of these works, remains to be seen. Which is why Gartner calls it the hype cycle, and generative AI is currently at the peak. Not that long ago, blockchain was supposed to solve many supply chain problems; now we are talking about AI.
Where inside supply chains AI will be applied is going to depend on where it is proven to be successful in actually solving the practical problems facing the practitioner. That means it has to go through the “trough of disillusionment” before viable solutions emerge that solve supply chain problems.
As the adage goes, the proof of the pudding lies in the eating.
G. 'Ravi’ Ravishankar is a faculty member at the Strategy, Entrepreneurship and Operations Division at the Leeds School of Business. He is a veteran of supply chain, lean transformation, implementing product innovation strategies and technology transfer from national laboratories.
His career has spanned a wide range of operating roles from president, CFO to engineering manager and director of innovation. He has worked in four continents on lean manufacturing, supply chain, logistics, product development, factory start-up, and business strategy. His industry experience includes, semiconductors, machinery, medical devices, food and beverage, chemicals, consulting and not-for-profit organizations.