A shortage of AI skills, high software costs and mounting data security fears top the list of hurdles facing SA manufacturers adopting AI, though most are still hopeful it will help them make more money.
A PwC study, AI in Operations: Revolutionising the Manufacturing Industry, covering over 400 manufacturing companies across EMEA (Europe, Middle East and Africa) found SA firms are eager adopters of AI.
“Nearly 70% of global respondents and 81% of SA respondents expect it to increase operating profits by at least three percentage points by 2030 and more than 56% of SA respondents anticipate an even greater rise of five percentage points in operating profits,” the report states.
Yet implementation lags.
“Though expectations are high, only 15% of SA respondents report significant financial returns from AI in operations,” said Pieter Theron, a partner and director at PwC SA.
“While 50% have moved beyond pilot phases, the pace of implementation and return on investment remain measured. However, this does not negate that AI is already delivering measurable improvements in decision-making, productivity, flexibility and delivery volumes,” he said.
Since the launch of ChatGPT there has been an enormous amount of attention on generative AI (Gen AI) as a way to increase efficiency across the entire enterprise, the report states.
Gen AI refers to AI models that can create new content — such as text, images, code, music or even product designs — rather than just analysing existing data.
PwC’s 28th Annual Global CEO Survey, Sub-Saharan Africa perspective reported that companies are seeing notable gains in efficiency, with 56% reporting increased employee productivity and 53% noting improvements in executive time management — both comparable to, or exceeding, global benchmarks.
AI can help increase sales and delivery volumes, improve decision-making, productivity, flexibility and reduce costs.
“Profitability is already being positively affected,” the research found.
“SA respondents report that it has increased sales and delivery volumes (57%), while 54% reported an increase in product prices. Perhaps even more impressive: nearly half of these companies (46%) are benefiting from AI. They are already creating new revenue from additional products and services (46%).”
According to the report, AI is helping reduce costs. Across the sample, the cost impact is most significant on operational areas, including energy, buildings, administration and personnel costs.
In several industrial sectors, a growing number of experienced workers are reaching retirement age, raising the risk of critical skills gaps.
AI offers a potential buffer by reducing the overall demand for labour in some areas, while also supporting front-line workers directly.
By guiding less-experienced staff through complex tasks — like visual quality checks — AI can help bridge the skills gap without compromising performance.
The automotive sector is emerging as the front-runner in AI maturity within manufacturing, the report notes, with computer-aided engineering tools like crash-test simulations becoming industry standard.
Use cases — a real-world application of AI to perform a task or improve a process — such as predictive maintenance, visual quality control, energy efficiency optimisation and detailed production scheduling are already in widespread use.
Pharmaceutical, life sciences and medical technology companies are leading the charge in research & development.
These sectors are already applying AI in product design optimisation, drug discovery, and regulatory compliance. GenAI tools such as Microsoft’s “AI for Science” platform are helping firms navigate complex clinical testing regimes.
By contrast, adoption in aerospace, defence and retail is markedly slower; however, the study acknowledges that this may be a result of a greater emphasis on using AI in areas such as marketing and customer service, rather than the operational areas that are the focus of this study.
Concerns about data loom large. Globally, the biggest barrier to implementation is poor data quality.
In SA, the lack of AI specialists and the high cost of AI tools are cited most frequently, followed by data privacy and cybersecurity concerns.
Many manufacturers operate fragmented IT and operational technology systems, with disparate, unstructured data formats creating additional obstacles.
To mitigate this, manufacturers are increasingly adopting hybrid solutions combining cloud and edge computing.
A promising new development is the rise of small language models (SLMs), which can operate on local infrastructure and offer an alternative to cloud-heavy large language models (LLMs).
Would you like to comment on this article?
Sign up (it's quick and free) or sign in now.
Please read our Comment Policy before commenting.