At its simplest level, the core purpose of the supply chain is to get the right amount of ‘stuff’ from one place to another at the right time, designed to transform raw materials into finished products for consumption as efficiently as possible. But while the theory may be simple enough, the reality is clearly far from it and is getting more complicated with every day that passes. Linear, localised supply chains with limited fluctuation in demand have been gradually displaced by complex, global, multimodal ones that hinge on just-in-time delivery and scramble to meet the heightening demands of consumers.
The businesses throughout the supply chain have of course taken steps to manage this complexity. RFID tags, for example, have been widely deployed to track goods as they weave their way across the globe, while manufacturers have adopted the latest technologies to churn out goods at a faster rate. However, a more fundamental issue faces the industry that has yet to be properly addressed: that there are no guarantees that supply will match demand.
The problem is that for decades businesses across the supply chain have relied upon manual processes to predict demand for their products and have made adjustments to their production and distribution lines accordingly. These decisions have often been based on previous track records, assumptions, and existing agreements with wholesalers.
However, with limited intelligence to back up these assumptions, these ‘guesstimates’ are essentially built on sand, resulting in two of the biggest problems facing the industry – overstocking and understocking. The former leads to perishable items going to waste, with all of the environmental consequences that entails, or inventory taking up valuable space in shops and warehouses, eating into margins; the latter results in lost sales and a reduction in customer satisfaction. This is not a trivial issue: according to data from IHL Group, a research firm, in 2015 the cost of understocking and overstocking to companies worldwide was $630billion and $470billion respectively.
Fortunately, recent developments in machine learning and artificial intelligence (AI) hold the key to this problem, with the technology able to deliver better decisions to businesses throughout their extended supply chains, unlocking the full business impact of intelligent data. In fact, according to a recent report from McKinsey, the total economic value to companies using AI across the supply chain could be between $1.3trillion and $2trillion a year.
How, then, could AI help to transform the supply chain and revolutionise its approach to matching supply with demand?
An intelligent data-driven approach to demand
The retail sector, as the consumer-facing end of the supply chain, is especially vulnerable to the changing whims of consumers, and typically bears the brunt of overstocking and understocking. However, we are already seeing major retailers adopt new AI-driven approaches to planning and replenishment.
Retailers are sitting on mountains of intelligence that typically goes unused or underused due to the sheer volume of data available and their inability to process it. By leveraging AI, retailers can make sense of this data, by analysing vast quantities of retailers’ internal data such as sales patterns and customer footfall, as well as weather forecasts and public holidays. AI solutions can then deliver granular predictions of customer demand across every product and in every store, to ensure that customers can always find the right products at the best prices.
However, while insights are useful, it is decisions that really deliver value to retailers. AI can make decisions on behalf of retailers to eliminate the burden of manual intervention and completely automate the replenishment and pricing process, reducing out of stock rates and improving sales, enabling retailers to devote more of their staffing resources to improving the customer experience.
Morrison’s, one of the UK’s largest grocery chains, is a prime example of this in action. The supermarket chain has changed its processes by using AI to improve demand planning and reinvigorate replenishment based on customer behaviour in every store. Focusing initially on ambient, frozen and long-life produce, then fresh food, it has increased profitability by delivering up to a 30% reduction in shelf gaps, cutting missed sales and eliminating waste. This is a powerful demonstration of how AI can play a fundamental role in enabling retailers to automate key processes and focus on delivering a better experience for the customer.
Moving up the chain
However, while it is encouraging to see decision-makers in the retail sector taking steps to invest in AI and optimise their processes, retail represents the very end of the supply chain. While retailers can finely tune their ordering and replenishment, there is a risk that the problem will simply be pushed upstream. Farmers will still grow more than they need to, which might just be left to rot, and manufacturers will still expend resources and effort on products that may never get sold.
To eliminate this waste and reduce inefficiency, businesses across the entire supply chain must adopt innovative technology to make best use of their data. When deployed in supply chain businesses, AI solutions could take decisions based on the analysis of vast reams of data to eliminate waste and ensure that only the necessary quantity of goods is produced, no more and no less. For example, if the data suggests that demand will be lower for a certain product, an AI solution could adjust production processes to ensure that a manufacturing facility doesn’t generate excess stock.
While some businesses may currently find it challenging to extract value from their data, given its volume and complexity, AI solutions can make hundreds of millions of calculations daily, far beyond the capacity of any human and deliver better decisions. In addition, existing legacy systems should not stop these supply chain businesses from applying AI to their core processes. Using cloud-based platforms, the most advanced AI solutions can sit comfortably on top of legacy technology, reducing capital expenditure and the time taken to scale up the solution.
Supply chains are undergoing a major digital transformation, and those organisations that want to remain competitive must extract the full value from their data. Advanced AI and machine learning will be critical to this, generating better insights and decisions for businesses, enabling them to reduce their waste and optimise their operations, creating a more efficient global supply chain.