Hi Guys,
There’s no question: Generative AI (GenAI) has captured the imagination - and attention - of business leaders across South Africa. What started as a series of pilots is now becoming core to how companies operate, serve customers, and make decisions. But as many businesses are discovering, the real magic doesn’t just come from having a Large Language Model (LLM) in your tech stack. It comes from making GenAI generate tangible ROI.
And that’s where a new evolution is taking shape: Retrieval-Augmented Generation (RAG), a transformative approach that connects AI systems to an organisation’s unique data, which allows these systems to generate responses that are both more accurate and more contextually relevant.
What RAG really does and why it matters
While LLMs have shown remarkable fluency in generating text, they are limited by the static nature of their training data. In fast-changing industries, this leads to outdated or misleading information. For businesses operating under strict regulatory frameworks, these shortcomings present real risks.
This is where RAG comes in. And for those of us leading transformation in complex, fast-moving environments, it’s the upgrade we need.
RAG makes additional data available to LLMs, increasing their efficiency and scalability without having to retrain the model. Businesses can expand their AI deployments more effectively and adjust as their needs evolve. Additionally, RAG’s ability to draw from a diverse set of external data sources means it can adapt to varying needs and applications, scaling the model to reach new use cases.
The result? Answers that aren’t just intelligent - they’re grounded in current, verifiable data that also improves the model’s performance on enterprise-specific queries. It’s no surprise that the RAG market is expected to grow from $1.2 billion in 2024 to more than $67 billion by 2034, at a compound annual growth rate (CAGR) of nearly 50%
(Precedence Research).
A new frontier: Agentic RAG
Now take that one step further. Imagine if your RAG-powered GenAI assistant could also plan tasks, make decisions, and act on your behalf. Unlike traditional AI systems that rely heavily on human prompts or structured inputs, Agentic AI is a software system that uses artificial intelligence to autonomously make decisions and take actions to achieve a set of objectives identified by humans who serve as orchestrators or managers. These systems can autonomously negotiate schedules, manage supply chains, respond to customers in natural language, and even execute operational tasks that meet the directed goal.
Crucially, agentic systems are not just reactive - they are proactive and collaborative, operating in increasingly human-like ways. This evolution opens doors to greater organisational agility, real-time decision-making, and massive efficiency gains. It also brings forth ethical questions and cultural shifts that must be addressed through thoughtful leadership. At Dell, we welcome collaborative efforts to set principles and standards for responsible AI development that build trust, safety and enable global harmonised standards and innovation.
 |
| Tony Bartlett, Director of Data Centre Compute |
A strategic playbook for South African CXOs
South African enterprises are embracing GenAI at an unprecedented pace. In the South African GenAI Roadmap 2025, a study published by World Wide Worx, Dell Technologies and Intel, 66% of respondents reported adoption of GenAI, up from 45% in 2024. Additionally, according to
The Institute of Risk Management South Africa, the local GenAI market is forecast to increase from $27.1 million in 2024 to $173.5 million by 2030.
Despite this momentum, challenges remain. The same study found that 40% of South African respondents identify high implementation costs as the top barrier to adopting AI, followed by data privacy concerns (30%) and lack of skilled talent (29%).
To prepare for this shift and capitalise on RAG/Agentic RAG, here are five steps CXOs can take:
1. Build the right data and infrastructure backbone
RAG systems, specifically Agentic RAG, are only as powerful as the data they’re fed. Many South African enterprises still struggle with fragmented data systems and poor data governance. Investing in cloud-native data architectures, real-time data pipelines, and standardised governance protocols is foundational. Embedding synthetic data generation and privacy-aware AI training mechanisms will also help accelerate scale while staying compliant with regional regulations.
2. Start building AI governance now
Agentic RAG systems operate with higher levels of autonomy, which raises concerns around accountability. Therefore, building robust AI governance frameworks is critical - these must include ethical guidelines, human oversight, audit trails, and scenario-based testing. Transparency will also be essential for businesses. As regulatory frameworks around AI evolve and the strategic importance of GenAI grows, South African organisations without clear governance ownership – or those relying solely on operational IT leadership – may face increased pressure to formalise their oversight mechanisms.
3. Drive workforce readiness
The future of work isn’t human or machine - it’s a human and machine collaboration. Equip your teams with the training, tools, and trust they need to work alongside AI systems. You’ll need AI translators, prompt engineers, and most of all, people who can see opportunities and act on them.
According to statista, the demand for AI and machine learning skills in South Arica is experiencing significant growth, with a projected annual growth rate of 26.22%. Businesses that invest in learning will be the ones that thrive.
4. Run focused pilots
You don’t need to deploy Agentic RAG enterprise-wide on day one. Adopt a “test-and-learn” operating model - run controlled pilots with clear KPIs based on business priorities, rapidly iterate based on insights, and scale what works - like onboarding documents, customer FAQs, or internal policy queries. Build momentum with wins that matter.
5. Connect AI to business value
Always tie your AI strategy back to business outcomes. Whether it’s faster decisions, better customer experiences, or more efficient teams, link GenAI adoption to measurable results. For example, a retailer in Germany uses RAG to power shopping assistants that check inventory in real time and deliver tailored product suggestions. The lesson? These technologies work best when they’re solving real business problems for people.
Final Thoughts: It’s Time to Evolve
For the South African CXO, the mandate is clear: don’t stop at LLMs. Lead your organisation into the next chapter of enterprise AI - one where intelligence is not just generated, but grounded, goal-driven, and trusted. As GenAI matures, the shift isn’t just about better models - it’s about better business outcomes. Smarter automation. Faster decisions. Deeper insights. Use it to build a business that’s not only more efficient but also more intelligent, adaptable, and ready for the future.
Article by Tony Bartlett, Director of Data Centre Compute, Dell Technologies South Africa
Aum Namah Shivay!
Subscribe to my
YOUTUBE Channel for historical tech, new gadgets, stationery, product reviews
Your Lifestyle and Technology Blog for the Last 12 Years!
