Supply Chain Optimization
Traditionally, enterprises run plans weekly or monthly to decide what they will ship, manufacture, and procure. They may drive these plans either from forecasts or from actual orders – or from both depending on lead times. But often, things change — machines fail, trucks break down, ships sink, factories flood, goods are stolen, yields are off, and weather continuously intercedes. The supply chain of manufacturers, distributors, retailers, and e-commerce companies is rife with uncertainty.
One example of a predictive application involves a retailer using inexpensive radio tags on SKUs to track the geographic location of inventory. Imagine a supply-chain “crystal ball” that was able predict when an expected order would be late or have fewer goods than expected. With this “crystal ball”, they can perform a “what-if” analysis on each potential late-order, for example, to find out what shortages would occur and what downstream consumers would be affected.
Here’s how: The predictive application predicts how late an order will likely be, and then simulates moving the order’s delivery date to the later time. As a result, consumers of the original order will have much fewer items in inventory, possibly resulting in shortages. The predictive application’s dashboard calls out the possible shortages as well as highlights the consumers of those items providing a great early-warning system for planners. With it they can plan around these anticipated events by shipping goods from other locations, or at least warn consumers downstream so they are not surprised. Planners could use the predictive application regularly, throughout the entire supply-chain, optimizing for anticipated events.
Cancer Clinical Trial Advisor
One healthcare solution company built a clinical trial advisor application for oncologists. The predictive application learned to predict the conditions under which a new drug clinical trial may be recommended. It used lab tests, health data, histories, and even genomic information to match patients to trials. It’s no longer possible for doctors to read every article detailing drug trials in every journal, so with an advisor like this, the doctor can consider new options and discover what drugs they may have available to them, even before seeing the patient for the first time.
Another health-care company uses predictive applications to “surveil” Electronic Medical Records in real-time and monitor them for patterns that may indicate a code-blue event such as sepsis shock. By using machine learning on all examples, a model can be built to predict situations that are likely to eventually result in this potentially fatal condition. These models are constantly improved in the surveillance application. When the likelihood of a code-blue event is high, the application can inject clinical workflow such as dispatching a nurse to a patient’s room to check a wound.
In milliseconds, digital advertising systems and e-commerce systems need to decide what to show to customers. The best-in-class systems now use predictive applications to do this. The systems dynamically change their decisions in real-time based on the behavior of the customer. For example, based on what they do online or in other selling channels, an offer to a customer may radically change. Recommendations dynamically change depending on how recently they visited a web page of a certain type, or whether they opened an email, or if they purchased something in a store. As time goes on, recommendations get better and better based on reviewing and learning from previous recommendations. This closed-loop learning process learns to predict in real-time customer interactions.
The Internet-of-Things (IoT) goes beyond the popular consumer devices like watches, goggles, light bulbs, electric outlets, and thermostats. Now entire networks of engineered industrial systems are outfitted with “smart” sensors that continuously report data on the health of the individual components and together can be used to predict the health of the entire system. New predictive applications are emerging in companies like those operating oil rigs, electrical grids, telecommunication systems, and security systems to plan preventive maintenance. It is very costly for any of these “network” companies to experience an outage. Customer satisfaction and revenues and are directly impacted. Being able to predict failures and proactively plan repairs to equipment before failure is invaluable. These companies use predictive applications to proactively dispatch service personnel to fix systems before they break, avoiding the deleterious effects.