Wanting to begin AI operations is not enough – action is needed
Artificial intelligence has become a corporate obsession. From the explosion of ChatGPT to the everyday integration of tools like Microsoft Copilot, companies echo with conversations about how AI is a competitive advantage or how it will improve productivity and customer experiences. According to Gartner, 79% of companies perceive AI as an essential part of their success story over the next two years, and yet only 20% are using it every day.
Enthusiasm isn’t the same as adoption. Companies are stalling at the starting line because they’re caught between the fear of falling behind and the confusion of how to begin. And they do need to begin because AI is stretching into every part of the business, from chatbots to email to documents to solutions.
The global AI market is forecast to reach R6.7 trillion by the end of 2027 at a compound annual growth rate of 36.2%, says MarketsandMarkets. It is adding value to multiple layers within the business that include workflows, robotics, IoT, back-office functionalities and customer experiences. Despite these numbers, uptake remains uneven, and in South Africa, many companies are wary of pouring money into ambitious AI projects that may be out-innovated within the year or that do not have a clear return. Companies are concerned that their data isn’t ready, their teams don’t have the right skills and what governance models they need to keep them compliant.
Implementation angst
There are three themes that are emerging in most conversations about AI. The first is the cost and the uncertainty of ROI. AI is perceived as expensive, requiring vast data lakes and infrastructure before results are possible. This misconception prevents leaders from experimenting with AI in smaller and more practical ways, particularly in areas of the business that could benefit from incremental AI innovation.
These concerns come with worries about data quality. Without clean and reliable data, AI deployments run the risk of failing before they start, and companies often underestimate just how much preparation is needed to build strong data foundations. Wedded to the data is security and governance. Leaders are legitimately cautious about where their data has to go and want to feel confident about how it is used and where it is stored.
These three challenges combined are leading to AI paralysis – companies know they can’t ignore AI, but they don’t know how to move forward without risking wasted spend or making regulatory missteps. Part of the issue is the speed at which AI is changing. Every month, new tools and models appear, and each one out-innovates the other. Analyst firm IDC has characterised the AI ecosystem as ‘simultaneously maturing and fragmenting’ because while tools are advancing rapidly, they are also multiplying, which makes it hard for companies to choose the solutions best suited to their needs.
Text | Chris Badenhorst
Photography | Stokkete
Chris Badenhorst is Head of Azure Core, Data and AI Services at Braintree. For more information, go to braintree.co.za.
