Artificial intelligence is reshaping how South African businesses think about software, skills and long term strategy. In a recent AI Impact Talk conversation, Riselle Rawthee, an AI research and development engineer at BBD Software, explained the real difference between the tools people use every day and the systems that sit behind them.
She drew a clear line between generative AI and agentic AI. Generative models create content and are often used to deliver structured results for specific tasks. Agentic systems use the same underlying models but are designed to act with more independence and critical reasoning. They can manage more complex processes because they are trained and prompted in a way that supports deeper thinking and more autonomy.
Many businesses see impressive public models and assume the technology works the same inside their organisations. Riselle stressed that quality still comes down to the data. The old rule of garbage in, garbage out is still true. Prototypes often work well because the data is clean and the scope is small. Real enterprise work involves large messy datasets that require care before any AI can produce reliable results.
She explained that businesses no longer need to build their own large models because of the heavy costs and expertise required. Instead, companies like BBD focus on leveraging established models such as GPT and Claude to create custom solutions that solve real problems. Chatbots are a simple example of how these tools already show up in everyday customer interactions.
For engineers, the shift is significant. The role is becoming less about writing every line of code and more about problem solving, understanding business needs and guiding how AI should be used. Coding agents can speed up work, but they cannot replace human judgment or collaboration. This has created a stronger emphasis on soft skills, teamwork and the ability to translate business requirements into workable solutions.
Riselle encouraged young engineers to learn through experimentation and to develop strong prompt engineering habits. She also advised both engineers and business leaders to stay curious, stay informed and understand that the best solution is rarely the most expensive one. It is the one that fits the problem, the data and the long term goals of the organisation.
