The global AI in supply chain market is projected to reach $9.94bn in 2025, predicts Precedence Research, and $192.5bn by 2034, at a CAGR of 39%. Automation, real-time monitoring, and demand forecasting are some of the biggest drivers of this growth, with AI-powered robotics, risk mitigation, hyper-personalisation tools and digital twins being used in supply chain optimisation.
Organisations are also investigating the potential of AI to improve distribution times and manage delivery timelines, with North America currently leading in adoption and innovation. In a study published by Frontiers, AI in the supply chain can lead to reductions of up to 50% in forecasting errors and 30% in operational costs, while potentially introducing a 20% decrease in inventory holdings and a 40% improvement in warehouse operations and labour management.
In South Africa, AI adoption isn’t lagging behind. A study published by the Durban University of Technology in the South African Journal of Economic and Management Sciences found that companies are very much aware of the value that AI brings, and how it can improve productivity, quality control and logistics. However, the study also points out that skills and workforce education are important in ensuring that these solutions deliver results. It is only one of the challenges inhibiting uptake in the country.
You need clean, standardised data or you’re going to get terrible results.
Stephan Mare, Maluti
A Wits university research report on the adoption of AI in the supply chain industry found that low digital readiness, skills shortages and high unemployment and limited infrastructure are contributing factors.
Despite the 2025 Budget promising R402bn to road and rail services and logistics, local infrastructure is still decaying. Many companies, especially those outside major cities, lack access to connectivity, which affects real-time data collection and the deployment of AI solutions that rely on cloud-based analytics and data streams. Fragmented logistics networks result in frequent congestion and bottlenecks that AI alone cannot solve. Structural weaknesses are consistently increasing costs and delaying the benefits that AI can deliver.
Over the past year, freight payloads decreased by 4.2% in January 2025 compared to January 2024, and freight transportation income dropped by 2.4% year-on-year for the same period, according to Stats SA. Major ports are experiencing severe congestion, with average vessel waiting times at the Durban port sitting at five days in April 2025. In Cape Town, peak waiting times are now more than 10 days in June 2025.
“We are talking with a few industry players at the moment on how to reduce the congestion, specifically at the ports, and make best use of the infrastructure,” says Michael Olivier, managing executive, OKgo, part of the Tracker Group. “The idea is to provide visibility into containers inbound and assess where the congestion comes from and what industries are most affected. The citrus industry, for example, has lost billions because of it.”
The idea is to provide visibility into containers inbound and assess where the congestion comes from and what industries are most affected. The citrus industry, for example, has lost billions because of it.
Michael Olivier, OKgo
Inefficient logistics cost the sector R5.27bn in 2024 thanks to unreliable schedules, deteriorating road and rail infrastructure and, of course, the port delays.
“Companies need visibility and data that allows them to predict and understand lead times so they can better plan for peak times, events and wastage,” says Olivier. “AI allows companies to get better predictions that reduces risk with more accurate lead times and can help identify delays in real-time, which can then improve decision-making around routes.”
While they refused to give comment, Massmart, Woolworths and Shoprite have all implemented customised logistics solutions that are allowing them to optimise their logistics. Massmart has invested in the Trackmatic portal to manage supply during peak events; Shoprite is using AI and ML to manage order accuracy and quality, specifically around fresh produce such as cheese, eggs, milk and poultry. The company has said that it has seen increased year-on-year sales as a result of the system as it has allowed for predictive stock planning and reduced customer wait times.
AI can also be used to gain visibility into stock temperature control. “You can log in and see if your container is below a certain temperature threshold and make adjustments if this changes at any point in the journey,” says OKgo’s Olivier. For Marius Jooste, general manager, Syntell, investing in AI-driven software capable of optimising back-end processes and leveraging the company’s data was essential. “We needed to get inside the data, and the machine learning and AI-driven intelligence offered by Maluti, the solution we implemented, has meant we’ve been able to improve outcomes for both our business and our clients,” he says. “We needed a solution that would allow us to ingest our data, understand the trends, improve predictions, and help us take actions based on the insights that were generated.”
Over the past six months, the company has seen a measurable shift in optimisation and system improvements, but, as Jooste says, there are still constraints in the system’s capabilities because customers are not as far along in the digitisation process. “Unfortunately, adoption is slow and our clients are taking longer to get on board, but for us, AI will always play a role and we will continue to test and improve adoption.”
Many companies are still trying to find their digital feet in South Africa, and they lack the data foundation to implement AI. Larger companies like Woolworths and Shoprite have the resources and infrastructure to introduce AI with relative ease, but for many companies, their data isn’t in a position to meet AI’s demands.
“You need clean, standardised data or you’re going to get terrible results,” says Stephan Mare, CEO, Maluti. “Once you have the data asset in place, then you can use your machine learning or AI models on top. If you want to start your AI journey and you don’t have solid data, you can forget about it.”
As companies collect more data, accuracy will increase. It also allows them to build digital twins that can revolutionise predictive planning and delays.
“Ultimately, we want to get to the point where we can use real-time data to be as accurate as the AI predicting how the wind blowing in Cape Town will impact congestion and suggesting alternative routes that will minimise the risk,” says Olivier. “Digital twins within the supply chain are where the AI is going to go, improving real-time visibility and prediction capabilities that change how companies manage their stock and logistics.”
South Africa’s logistics infrastructure is attracting investment and it is undergoing modernisation, but the effects will take time to materialise. Until then, AI may be one of the best ways to minimise the impact of delays and congestion on the supply chain in the country.
* Article first published on brainstorm.itweb.co.za
Share