AI is set to have a profound impact on society, perhaps more potent than any technology humanity has ever witnessed. It provokes fear and joy in equal measure, but while some may worry about its implications for society and the future of employment, its benefits for business are undeniable. And these are already starting to be realized. In a recent survey by Accenture, 70% of executives said they are significantly increasing investments in AI compared with two years ago.
Improvements in AI have been driven by machine learning. Machine learning is a form of artificial intelligence able to learn without a human programming it to specifically find something. Machine learning algorithms do this by searching large data sets for meaningful patterns, from which future events can be predicted or classified. It finds the sort of patterns that are often imperceptible to traditional statistical techniques because of their apparently random nature.
One sector already realizing the benefits of machine learning is supply chain, where we could eventually see end-to-end automation. In Gartner’s recent ‘Predicts 2016: Reimagine SCP Capabilities to Survive,’ the research firm said supply chain organizations expected the level of machine automation in their supply chain processes to double in the next five years, and given the pace of change already, it is likely this estimate is conservative.
Its uses in the supply chain are many, from manufacturing to logistics. Its most important function, though, is in reducing risk, which it does in a variety of ways.
Firstly, machine learning will remove risks to employees by taking them out of more menial jobs, particularly those involving heavy machinery in which accidents are more likely to take place. For example, it is machine learning that makes driverless cars work. Just recently Apple confirmed what many suspected, that they in addition to Google, Amazon, and major car companies including Volvo, are also investing in this area, and once governments get serious about putting the infrastructure in place for driverless cars, it is only a matter of time before we see widespread adoption. Driverless cars will soon become driverless trucks, automating the transportation of goods. This solves a range of issues on the road, removing overly tired drivers, who are a tremendous hazard both to themselves and others, as well as reducing the chance of wrong routes being taken. Similarly, machine learning will decrease the need for warehouse workers. ARC did a survey on investment priorities in the warehouse this year and about 15% of warehouse executives said that procuring autonomous mobile robots in the next three years was a priority for them. This means not just less accidents in the workplace but far greater efficiencies around movement of products.
Perhaps the most important application next year in AI for supply chain managers will be in the ability of machine learning algorithms to analyze the wealth of data coming in from IoT and other sources. Unlike standard supply-management software, machine learning systems can collect, analyze, and adjust large data sets from a wide variety of sources efficiently and cheaply, combining every piece of data from internal factors such as staff sick days through to external sources such as weather data. This can then be used to build a complete picture of where potential stumbling blocks could arise and exact levels of demand to ensure that there is never an unnecessary surplus of stock. This is a particularly pressing issue for grocery. If you order too little, your customers will lose patience with empty shelves and go elsewhere. If you order too much, food will go off and you will be left with a huge amount of waste. According to a McKinsey report, early adopters of machine learning for replenishment in retail have already seen a number of benefits. These include reductions of up to 80% in stock-out rates, reductions of more than 10% in days of inventory on hand and write-offs, and gross margin growth of up to 9%.
One of the greatest risks to supply chains is the weather. Weather has an impact on everything from farming to transportation, causing spikes in demand for specific products that must be anticipated and supplied for and unplanned downtime that is extremely expensive. Heat waves and drought in Russia in 2010, for example, resulted in economic losses of $15 billion. In a recent survey of supply chain professionals by the UK Met Office, however, just 16% said use commercial weather data. The results they achieve are persuasive, with 57% of those that use paid-for data saying they had better sales forecast accuracy, 51% that they had better on-shelf availability, and 43% that they had reduced waste. Last year, IBM acquired The Weather Company to make the best use of this data, applying cognitive technology to it for optimal forecasting. Together, IBM and The Weather Company developed ‘Deep Thunder’, a hyper-local forecasting tool for business customers. Deep Thunder processes more than 100 terabytes of third-party weather and location data a day taken from hundreds of weather stations, combining this with real-time news information streamed from social feeds and news reports, as well as historical weather data. They then apply forecasting models and machine learning algorithms from its Watson program to the data sets to discover insights. These systems produce far more reliable weather forecasts, including the kind of location-specific information about impacts of storms, hurricanes and typhoons that is vital for supply chains to know. IBM sends the information to its own supply chain managers, as well as enterprises which they can leverage to better anticipate extreme weather as soon as possible and put in place strategies to minimize the impact. Bad weather can disrupt supply chains globally on many levels.
Machine learning not only identifies patterns that indicate risk, it can also recommend mitigating actions by examining previous responses to similar situations and identifying which were most successful. The rise of AI is enabling a future in which a supply chain will be a dynamic organism able to adapt rapidly to a constantly changing environment, identifying risks and offering solutions to deal with them in near real-time which will lead to a more efficient, safer, cost-effective supply chain.