Whereas e-commerce companies are digital natives -- that is, they were built from the ground up for the digital economy -- traditional manufacturing and consumer packaged goods (CPG) companies are hamstrung by the legacy systems, processes and tools that they’ve used for years. These systems prevent them from competing on the same level as digitally native companies.
While consumers increasingly expect online orders to be delivered overnight or as close to it, companies relying on the traditional supply chain are still hanging onto economic theories developed 30 or 40 years ago. Adding in the uncertainties of the Covid-19 economy, the differences between nimble e-commerce companies and their laggard traditional supply-chain counterparts have never been more pronounced.
Supply chain leaders aren’t blind to this issue; They understand this problem deeply, because their inability to react quickly enough to demand or overall market changes is impacting both their revenues and their brand image. Companies up and down the supply chain are undergoing a massive transformation, re-evaluating tools, systems, processes, people, and technology in order to be more competitive with their e-commerce brethren.
Transforming the supply chain
The primary challenge for these companies is that, unlike Amazon, Airbnb, or Apple, they cannot stock inventory electronically or virtually, which means they must find another way of gaining a competitive advantage. The future of competition is not a David and Goliath fight; Rather, it’s the Tortoise and the Hare story, with the hare eating the tortoise’s lunch. Agility, nimbleness, adaptability, and resilience are the keywords for the future, providing a better competitive advantage than traditional size and scale.
Even though supply chain companies recognize this challenge, many have invested millions — or even billions — of dollars in factories, transportation, and other infrastructure and can’t simply throw it all away and start from scratch. Instead, they need to leverage and transform their existing investments to allow for ultimate agility. In other words, they must change engines mid-air while in full flight. This is where AI can help.
How does AI fit into supply chain transformation?
Today, 90 to 95 percent of corporations in the supply chain are conducting AI/ML pilot projects, but many are just chasing the latest technology rather than thinking strategically about where to apply AI/ML techniques for the greatest impact.
AI is best deployed in situations where it can interpret what’s happening in the supply chain, make sense of it, learn from it, and then give humans the decision support they need. It’s through this so-called augmented decision-making that AI can really make a difference in supply chain transformation.
While there are certain situations where automating decisions may be possible and even make sense, those decisions have a low impact threshold. For those decisions that really impact supply chain transformation, human accountability and judgment must always be part of the decision making process.
So where should you look to AI/ML to truly transform your supply chain? Everywhere there is manual work. If you can automate those manual processes, you gain cost efficiencies.
There are three main use cases that are ripe for immediate AI impact:
- Demand forecasting
- A function that’s typically done by humans using traditional techniques, such as the 80/20 rule (in which the focus is on 20 percent of the SKUs that drive 80 percent of all revenue), demand forecasting is becoming increasingly difficult. The sheer volume of data available for every SKU (e.g., consumer sentiment) multiplies the number of variables that humans must process 100- or even 1,000-fold. This is where machine learning and AI provide tremendous value and efficiency, given the availability of vast compute power at an economical cost.
- Demand sensing
- Demand sensing takes the traditional history-driven demand-planning process and enhances it with real-time data from market events that influence demand patterns. While anticipating demand is never simple, it has only become increasingly difficult in the COVID-19 economy. Where do you anticipate demand coming from and when? What are the demand indicators or signals?
- When data volume and velocity change by the minute or hour, forecasters face a near-impossible task; AI can help make the task more manageable. Unlike humans, machines aren’t overwhelmed by data intensity and instead, can use the data to increase decision fidelity while keeping humans in the loop to arrive at a final signal.
- Supply and raw material planning/inventory
- Once you have the ability to establish the demand signal, however volatile it is, you must be agile enough to respond to it immediately. This is the key: the ability to respond to situations. You need to decide how much raw material inventory to hold, how much to move, how much to anticipate, and how much to pull-in, push-out, or even cancel. Your organization's ability to adapt and be agile becomes a competitive advantage over time.
So, yes, AI can help supply chain transformation because it enables companies to leverage a lot of intelligence in different use cases, but to be successful, AI must be embedded in — not added or bolted onto — a given process. Once AI proves successful end-to-end in the use cases I mentioned above, additional use cases, such as predictive maintenance, will clearly emerge.
While AI is not a silver bullet, it’s definitely a powerful capability in the supply chain toolbox. And when it’s embedded within a business process, results will follow.