With global food demand expected to rise by 70% by 2050, efficient management of the food supply chain is imperative to promote food security and minimise food waste. That’s according to a recent academic paper by Watson Munyanyi, an academic researcher at the University of Johannesburg, and David Pooe, director of the university’s school of management.

‘[However,] traditional methods of supply chain management often fall short due to challenges such as inaccurate demand forecasting, inefficiencies in inventory management and a lack of transparency,’ they write.

‘One of the most promising solutions is AI [artificial intelligence], where machines can learn through experience to adjust to new inputs and perform human-like tasks.

‘Through predictive analytics and optimisation algorithms, AI-driven systems can forecast demand, optimise transportation routes and minimise waste, ensuring that food reaches consumers efficiently and with minimal spoilage.’

The use case is obvious in South Africa, where, according to the Consumer Goods Council of South Africa, 31% of households experience hunger and another 30% are at risk of food insecurity – even as 10.3 million tons of food is being wasted each year.

Munyanyi and Pooe highlight several AI technologies, including machine learning (ML) and natural language processing (NLP), both of which play a critical role in predictive analytics used by a growing number of retailers.

‘[Machine learning’s] ability to analyse historical data allows ML algorithms to predict future demand with high accuracy. As such, companies can optimise inventory levels, avoiding stockouts and overstock situations.’

In addition, ML models ‘can identify the most efficient routes, transportation methods and logistics strategies, thereby reducing costs and improving delivery times’… and faster delivery times mean products with a short shelf life can reach consumers in a fresher state.

‘NLP [is] useful in supply chain management as it facilitates automated communication with suppliers, streamlining the ordering process and handling inquiries efficiently,’ they write.

‘Predictive analytics anticipate potential spoilage, redirecting products where needed most and ensuring only high-quality produce reaches consumers, thus reducing retail-level waste. AI-powered systems optimise distribution routes, decreasing transportation distances and fuel consumption.’

It is thanks to AI and predictive analytics that South African chicken restaurant chain Nando’s has been able to cut down its food waste by about 20%. As well as reducing wastage, by enabling Nando’s store managers to accurately order stock and prepare chickens, the technology helps prevent shortages and improve service speed.

‘QSRs [quick-service restaurants] rely on time-sensitive production planning. Food has a shelf life for a reason, quality is impacted, staff are impacted, and naturally, your customers are impacted,’ says Chris Swanepoel, Nando’s restaurant support director. ‘Running short of chicken at a chicken restaurant is undoubtedly the most stressful outcome for any manager to deal with, as are high levels of waste and idle staff,’ he says.

Nando’s employs a system, devised by global data analytics provider Predictive Technologies, that uses a combination of economics, data science, AI and behavioural insights. Predictive Technologies’ customers range from the restaurant and food processing sector to retail, hospitality and financial services.

‘Keeping high levels of stock is expensive, particularly if the product has a short shelf life. Our stock optimisation tool makes sure stock-keeping units end up at the branches where they are most likely to sell. It also provides early warning systems so that stores do not run out of products, and can automate stock ordering. This system also helps identify products which are likely to do well early on, and to cut unpopular products,’ it says.

Another of its South African clients is the Hungry Lion fast-food franchise, which uses data analytics to forecast demand and manage staff work schedules at its 230 stores in Southern Africa.

Hungry Lion reported seeing a 40% improvement in demand forecasting and a 14-percentage point reduction (from 34% to 20%) in the cost of staff scheduling errors.

‘No person can predict that accurately – but a machine algorithm can. It looks at the historic values, actual performance, as well as behaviour and economic insights to predict sales volumes,’ says Gideon Jacobs, head of productivity at Hungry Lion. ‘AI decreases the risk of branch managers attempting to predict sales, stock and staff volume, especially during turbulent times.’

The Shoprite Group has been using AI and ML for a number of years to predict sales at its stores. Sanjeev Raghubir, the group’s sustainability manager, explains that stock replenishment orders are placed automatically, to ensure that stock is always available for customers while simultaneously reducing food waste.

‘For example, a store close to the finish line of an annual sporting event will automatically be replenished with additional convenience meals for that single day of the year,’ he says.

One example of an automatic procure-to-pay ordering system is Cape Town-based Arch Retail’s eReplenish.

‘Complete integration is vital for successful retail automation,’ Casey-Lee Venter, marketing, communications and brand head at Arch Retail, says in an interview with online magazine Supermarket & Retailer.

‘Our eReplenish system doesn’t just track inventory; it integrates with suppliers, manages purchase orders, tracks deliveries and handles invoice processing. This end-to-end automation eliminates manual errors and ensures optimal stock levels.’

It can also adjust its order volumes to match consumer demands and the retailer’s cashflow. So instead of ordering large quantities once a month – which could conceivably lead to food waste and unnecessary expense – it can make smaller, more frequent orders based on real-time consumption patterns.

‘While AI might be considered a buzzword by some, solutions such as our analytics tools are making it real and accessible for SA’s retailers,’ says Venter. ‘The technology is mature enough to deliver measurable benefits, but it requires thoughtful implementation and ongoing optimisation,’ she says.

Aside from reducing food waste, retailers are also using AI to improve efficiencies in other areas, for example, to improve the customer experience.

Kerry Janse van Rensburg is the person driving the seamless blending of digital technology into Pick n Pay’s operations – online and in-store – while also integrating its customer loyalty programme – Smart Shopper – into the customer’s shopping experience.

‘The Smart Shopper data is gold for us,’ Janse van Rensburg tells Daily Maverick. ‘It is that unique identifier between online and offline.’

An understanding of the data leads to a greater understanding of customer habits, and then delivering what they want, where they are.

‘We are giving [customers] personalised offers and hyper-personalised messages and being in the moment with them, which AI will help us do,’ she says.

Janse van Rensburg expects that digital technologies will completely redefine grocery shopping. ‘AI will help predict what customers need before they search, personalising offers in real time. Data analytics will drive smarter decisions across pricing, stock and promotions,’ she says.

As well as improved demand forecasting and enhanced inventory management, academics Munyanyi and Pooe also identify AI’s ability to improve transparency and traceability of products, which would help improve food security. In addition, ‘AI-powered systems optimise distribution routes, decreasing transportation distances and fuel consumption’.

They do acknowledge, though, that there are challenges in employing AI in Africa – technological integration, data privacy concerns and the need for substantial investment in infrastructure and skills development.

‘Collaboration among stakeholders – farmers, processors, distributors, retailers and technology providers – is crucial for successful AI integration, enabling cost-sharing and collective progress,’ they write.

Images: Unsplash, iStock