Artificial intelligence promises a transformation on par with the proliferation of the internet or mobile technology, especially across the diverse economies of Southern Africa. Yet for the region’s small and mid-sized enterprises (SMEs), the road to responsible, sustainable AI remains as challenging as it is promising. What’s increasingly clear is that success depends as much on foundational practices like data quality and robust access controls as it does on cutting-edge models and algorithms.
Building a Centre of Excellence: Why It Matters
For enterprises large and small, establishing an AI Centre of Excellence (CoE) has become the touchstone of safe, scalable innovation. While formal CoEs remain relatively rare among SMEs, Southern Africa is witnessing a growing trend: sector and regional centres supporting businesses through shared expertise and governance frameworks. These CoEs don’t just centralize technical skills; they create the discipline needed to scale AI without sacrificing trust or compliance.
Latest Data on AI Adoption Among South African SMEs in 2025
South African SME adoption of AI continues to accelerate through late 2024 and into 2025, with notable gains in both the rate of uptake and depth of integration. Enthusiasm and ROI remain among the main motivators for adoption, especially in process automation and analytics. However persistent barriers such as skills shortages, data readiness, and funding constraints continue to limit widespread adoption. Below is an updated analysis, reflecting the latest available market data and trends as of 2025.
Adoption Rate:
Estimates now show that around 56–60% of South African SMEs have begun integrating AI-powered tools into at least one aspect of operations (up from <50% in early 2024).
Full integration into daily processes is observed in roughly 30–35% of SMEs.
ROI & Use Cases:
Operational efficiency improvements remain prominent, with some SMEs reporting up to 35% gains, especially in sectors like retail, logistics, and financial services.
Most common implementations:
Chatbots/virtual assistants for customer service (∼45% of AI adopters)
Predictive analytics and business intelligence (∼40%)
Workflow/process automation (∼35%)
Personalisation engines for marketing/sales (∼20%)
Barriers to Broader Adoption:
Skills gap is the most-cited hurdle; more than 60% of non-adopting SMEs highlight limited AI and data science expertise.
Data infrastructure is frequently underdeveloped (∼55% of respondents).
Resource/financial constraints are reported by over half of non-adopters.
Concerns about change management and cybersecurity are increasingly referenced.
Data Quality as a Dealbreaker
No matter the size of the business, one factor routinely separates AI leaders from laggards: the quality of their data. For SMEs just embarking on this journey, it’s tempting to focus purely on tooling and model sophistication. But research and real-world case studies make it plain: over 80% of failed AI initiatives point to poor data quality as the root cause. In my experience at Keyrus South Africa, successful AI CoEs don’t leave data cleanliness to chance. They establish clear definitions for data, invest in careful cleansing, and embed ongoing quality checks within every project lifecycle even when working with lean teams.
The Crucial Role of Access Control
As more business functions find value in AI, the need for structured data governance escalates. Role-Based Access Control (RBAC), long a mainstay in larger organisations, is increasingly vital for SMEs as well. RBAC helps define exactly who gets access to which data and for what purpose, an essential guardrail given the sensitivity of business and customer information.
The reality, however, is that few SMEs get this right from the outset. Studies from 2023–2024 highlight that adoption of RBAC and more advanced access management controls remains patchy across smaller firms, most often due to resource and awareness challenges. Yet, the consequences of missteps such as unintended data exposure or regulatory breaches can be existential for SMEs.
Regional Trends and Supports
There’s cause for optimism. Most Southern African governments and industry groups now recognise the SME gap in AI capability and are working to close it. Shared CoEs, national policy initiatives, and business incubators are helping bridge technical and regulatory divides. Practical success stories like SME clusters in financial services or agri-tech leveraging community-led Centres of Excellence demonstrate that progress is not only possible, but accelerating where support structures exist.
Practical Advice for SME Leaders
Start small, but start right. Even one dedicated data steward and basic access rules go a long way.
Demand clarity in your data. Standardise, document, and clean before building.
Seek out or join collaborative CoEs, whether through sector bodies, universities, or local networks.
Upskill your team continuously, prioritising responsible data handling and governance.
Scaling Trust, Not Just Technology
Amid the rapid rise of AI, it’s easy to get distracted by hype and big promises. Experience across Keyrus South Africa and the region shows that, for SMEs, the path to safe scaling isn’t about massive investments or waiting for perfect conditions. It’s about disciplined foundations: robust data quality, strong access control, and a culture that values trustworthy innovation. Build on these, and a Centre of Excellence becomes not just a technical function but your organisation’s compass, guiding sustainable AI for years to come.
Ready to build trust into your AI journey? Connect with Keyrus to find out how we’re helping SMEs lay the groundwork for the future. Contact us at sales@keyrus.co.za.
References:
SACAIR 2024 Proceedings – Responsible and Ethical AI in South Africa:
https://2024.sacair.org.za/wp-content/uploads/2024/12/SACAIR24_vol_II.pdf
Ramaliba, T. "AI to Enhance Data Quality Management in the Banking Sector" (2024):
Oldemeyer, L. et al. "Investigation of Artificial Intelligence in SMEs: Systematic Review" (2024):
https://www.ssbfnet.com/ojs/index.php/ijrbs/article/view/3561
McNulty, N. "How Generative AI Can Transform South African SMMEs" (2024):
https://medium.com/@niall.mcnulty/how-generative-ai-can-transform-south-african-smes-fe05cdad22e5
Trust.org "AI Governance in South Africa" (2025):
https://www.trust.org/toolkit/part-2-emerging-ai-governance-in-africa/ai-governance-in-south-africa/
Nexia SAB&T "SMEs Embrace AI" (2024):
https://www.nexia-sabt.co.za/smes-embrace-ai-to-boost-efficiency-and-competitiveness/
South Africa National Artificial Intelligence Policy Framework (2024):