How AI for supply chain is transforming customer experience
Meagan Johnson Meagan Johnson
August 27 7 min

How AI for supply chain is transforming customer experience

The retail supply chain of today looks a whole lot different from the retail supply chain of twenty or even ten years ago. That’s because today it’s the customer who is sitting at the master switchboard, directing the traffic of a retailer’s supply chain. Retailers can no longer simply push products through to customers as they once did. The customer is now hyper-connected and in control, and this creates more variability, volatility and complexity than ever before.

Stores, warehouses, factories and logistics networks are adopting new methods of digitization to help them meet customers wherever they may be, when they want, with the right price and the right products. This hyper-connectivity creates a massive and growing abundance of data, and it’s not always easy to discern which signals are most relevant and actionable⁠—and which are just noise. Simply gaining visibility on data through a central control tower is not enough to adapt to and meet customers’ changing needs and behaviors.

What is needed is an active learning supply chain, with more dynamic and shorter planning and execution cycles. AI that continuously learns and improves brings this vision into the realm of the possible: AI can make sense of data and use it as the basis for recommended actions; it can analyze disparate data points to better manage and minimize inventory disruptions; and it can pluck all the different strings in the supply chain together in real-time orchestration, so that there are fewer and fewer customers who journey to a store, only to find themselves standing in front of an empty shelf where the product they want should be.

Data is crucial, but it’s not where the value lies

Many retailers have the data, some more than others, but the competitive advantage isn’t in the data itself (which is by no means scarce). Rather, the competitive advantage lies in the aggregation and interpretation of the right data, from multiple sources (and thus, across business functions), and in using this data as the basis to make faster, better-informed business decisions.

This is where AI gives retailers an edge. With machine learning software that analyzes and learns from data, it’s possible to move beyond dashboards, which provide users with visibility, to AI-recommended actions. These recommendations, which should be realistically executable across the relevant points of the supply chain, enable businesses to better keep pace with today’s hyper-connected consumers.

Take one example: a large company recently reported that their ERP system produced some 800 alerts per day, making it extremely difficult for employees to determine which should be acted upon and in what order. They also needed to understand what could realistically be addressed within a given execution time horizon and what the value at risk was if not acted upon. By using AI, the retailer showed that the number of actionable alerts was quite a bit smaller than the number they were receiving. What’s more, some real disruptions were never picked up by the deterministic rear-view-looking ERP system. This is simply an example to show that the ability to find patterns in data and convert only the relevant, high value, actionable insights into intelligence that can be executed at the right time, in a connected and synchronized manner across the supply chain, is what will really make a difference in the eyes of the customer.

Connecting and interpreting traditional data points such as inventory on hand, truck ETAs, production dates or more unstructured data such as that produced from cameras, wifi, beacons or sensors, is key to determining what information is most relevant in addressing high-value at-risk potential disruptions. If we think of an event like out of stock, which has long been considered one of the biggest pain points for retailers, the ability to discern which out-of-stock scenarios are the most important to address and which should be actioned upon when it matters for the customer, will allow both for decision-makers to spend more time on value-added activities and for execution- level roles, such as store associates, to make greater gains in productivity.

With AI, retailers can respond faster to disruptions in the supply chain

If we think of a given product and try to imagine the journey that it goes on across the customer-centric supply chain, there are potentially smaller batch production runs to accommodate more fragmented, customized demand, and there is most certainly more manipulation due to the various customer channels that it may move through or between. Therefore, in addition to noise, there is also an increased potential for data distortion.

Retailers can respond faster to disruptions in the supply chain, with AI.

For example, one retailer told us that around one-fifth of its e-comm orders are left unfulfilled due to incorrect stock-on-hand inventory data. Retailers need to be able to detect or predict where these kinds of distortions may exist or could arise, so that, if and when they happen, they are prepared to quickly reconcile discrepancies and ensure they are considered when prescribing a course of action or mitigation plan. AI has the potential to help retailers address some of the challenges of inventory disruption. Computer vision, for example, can be harnessed to better detect and count products in the front and back store and identify misplaced items or damaged goods, helping to maximize on-shelf availability.

From localized insights to big-picture orchestration

Detecting relevant signals and predicting where disruptions may occur in order to support decision-makers with prescriptive recommendations is already one step ahead of control-tower visibility. But none of this has much value if the execution isn’t properly orchestrated within the defined execution time horizon. With a more complex and interconnected supply chain, the standard planning-to-execution lag effect gets multiplied, and what was a carefully crafted plan is now less impactful or even irrelevant because the execution was not synchronized across the right nodes and actors in the supply chain.

From localized insights to big-picture orchestration.

Why does this matter for AI, and how can it help here? Execution happens at the store level and warehouse level and truck level and, arguably, even at the individual associate level. The only way to devise a good and actionable execution plan (or mitigation plan in some cases) is to be able to analyze and capture data and signals at those same levels. This implies collecting a high volume of data, which—unlike legacy systems and models—deep learning models can accommodate. Most importantly, over time AI can capture feedback or interpret signals at the ‘local’ level and continuously learn and be retrained to deliver richer prescriptive guidance and localized planning and execution. This also makes it easier to predict where data distortions may exist to more quickly put together a set of prioritized tasks, orchestrated across all relevant nodes and players, to fix the internal data problems before they become a customer problem.

Upon executing on the AI system’s predictions and recommendations, new data points can be captured that link the prescribed plan to the effectiveness of the actual execution. The insights from this data can become a point of differentiation as they can significantly reduce the time between planning and execution by creating a continuous and more real-time feedback loop. This allows for more dynamic planning and continuous improvement and directly supports the need for supply chain agility in this complex, customer-first landscape.

Delivering a seamless shopping experience in the age of the connected consumer

Today’s on-demand economy will require retailers, manufacturers and logistics companies to adopt innovative new approaches. AI has the ability to provide richer and more localized insights that can lead to more dynamic and optimal planning and execution. It connects disparate data sources and interprets new information, cutting through the noise and offering prescriptive solutions to guide employees on what is most relevant and actionable. Moreover, coupled with quality input data, AI is capable of identifying and predicting discrepancies, thereby minimizing inventory disruption. Finally, AI can orchestrate execution and capture new insights for more localized and dynamic planning across the whole supply chain. With AI that learns and improves, retailers can win the customers of today by giving them more of what they want: an easy, reliable and fun shopping experience, so that they never again leave the checkout empty handed.

Going to be at Groceryshop on September 15-19 at the Venetian in Las Vegas? Book a meeting with one of our experts to learn how AI can optimize your supply chain.