COVID-19: A Call for AI Powered Supply Chain Decision Making

Datamation content and product recommendations are editorially independent. We may make money when you click on links to our partners. Learn More .

“It was the best of times. It was the worst of times…”

Charles Dickens could have very well said this about COVID-19. The pandemic is causing some strange anomalies unlike any we have witnessed in our lifetimes. Companies across the board are experiencing unforeseen shocks to the system on the demand and supply side simultaneously.

For some, business is booming, but for many, these are challenging times. I will share three compelling stories of companies that are going through a significant shift in their business baseline and highlight how AI is helping them get through the challenges.

All of them had implemented best-in-class planning systems prior to COVID-19. However, the current crisis forced them to rethink their approaches and power their decisions with artificial intelligence.

Predicting shifting demand across product mix using external data

A leading beverage company is faced with unprecedented levels of volatility with its sales patterns. As the pandemic spread across countries and as stay-at-home orders went into effect, the company started experiencing significant demand shifts in its portfolio.

For example, the SKUs meant for the hospitality and restaurant industry declined drastically. But the consume-at-home part of the portfolio radically picked up, resulting in a breakpoint from the historic patterns. The forecasting approaches anchored upon leveraging historical shipments and promotions suddenly came up short, as they weren’t responsive enough to the dramatic shifts in demand.

The company responded by bringing in external data sources such as publicly availability Mobility indices, which track shifts in outdoor activity, and by leveraging AI, combined with internal data, to learn from those patterns. As cities begin to open up and away-from-home consumption picks up again, the company can now identify how increases in mobility correlate with demand shifts across the product mix. Incorporating external data into their forecasting process to generate more accurate forecasts yielded double-digit percentage improvements in forecast accuracy.

Optimizing the production plans in light of supply disruptions:

A major packaged food company was facing serious shortages in packaging material due to logistics disruptions. At the same time, the demand for the company’s products spiked because many families increased consumption as a result of stay-at-home policies. The company also faced the additional challenge of reduced capacity at production facilities due to employee safety and social distancing measures.

The simultaneous material and capacity constraints resulted in the company making some tough tradeoffs on what to make and what not to make. It had to not only consider what to produce to maximize profit but also factor in the adverse impact of too many production changeovers. The planning system they already had in place provided directional guidance with rough-cut capacity planning.

But by implementing optimization algorithms to complement the planning system, they could bring finer definition to the production mix while respecting the material and capacity constraints. The company decided to narrow down the assortment to more-profitable, faster-moving SKUs, resulting in reduced changeovers and maximized profit potential. The improved product availability helped in gaining an edge over their competition.

Ramping up online sales amidst store closures:

A major apparel retailer faced the challenge of nationwide store closures and all of its sales shifting to e-commerce. As the shutdowns occurred, new seasonal inventory was arriving from overseas and threated to overwhelm the current Distribution Center capacity, which was already holding inventory for their brick-and-mortar stores.

It was not a trivial task to turn off the store locations in their planning system and reallocate inventory to online sales and overflow DCs. Advanced algorithms were developed and deployed as apps and shared across the network of planners performing inventory allocation in a matter of two weeks! These algorithms consider complex business rules and help planners decide what promotions to run, when to use offsite storage, or on inventory liquidation across nearly 100,000 SKUs.

In each of the above instances, the existing planning systems that are meant for the normal course of business were significantly challenged to keep up with the seismic shifts resulting from COVID-19. In a world where “never normal” is the new normal, organizations need to rethink their decisioning.

When faced with extreme disruptions that are becoming increasingly common, there is an ever-increasing need to power planning processes with advanced algorithms that can learn, evolve and simultaneously consider complex tradeoffs. However, such algorithms don’t need to sit on islands as one-off efforts. Nor do they need to require major rip-and-replace.

Instead, emerging digital platforms allow these algorithms to be operationalized and democratized in the form of apps that can be deployed quickly at scale, embedding them in the midst of the existing processes that then become increasingly interconnected. That is digital transformation delivered in bite sizes – and it has the potential to transform even the new “never normal” into the best of times!

ABOUT THE AUTHOR:

Dr. Madhav Durbha is the Group Vice President of Industry Strategy at LLamasoft, where his team helps customers solve various supply chain challenges. Dr. Durbha received his Ph.D. in chemical engineering from the University of Florida and his bachelor’s degree in chemical engineering from the Indian Institute of Technology at Madras.