Technology

SA businesses embrace AI, but are they getting real value?

South African businesses are embracing artificial intelligence with speed, but a latest tudy suggests many are still struggling to turn that enthusiasm into measurable results.

The SA AI Index 2026, compiled by World Wide Worx with Altron Digital Business, found that 93% of large organisations now use AI. That is more than double the 45% recorded in 2024, and up from 67% last year.

Yet just 51% of organisations using AI say it has a positive effect on productivity. The figure was 96% in 2024 and 83% in 2025. In plain terms, AI has become common in the workplace, but confidence that it is making work better has fallen sharply.

The gap is not about buying software or giving staff access to chatbots. The index measures how deeply AI is built into business operations, plus deployment, workforce adoption, governance and the ability to measure value. The average organisation scored 34.3 out of 100. The value category was weakest, at 16.6.

Many companies are still testing tools rather than integrating them into everyday processes. More than half, 52%, remain at the experimental stage. And 73% have no tools to assess whether AI is delivering value, while 71% cannot quantify the cost savings.

The difference shows up in productivity figures. Among organisations with AI running in live business processes, 80% report a positive productivity impact. That falls to 44% among those experimenting, and just 6% among those using AI only occasionally.

For workers, the findings point to a period of adjustment rather than an instant revolution. AI may help with routine tasks, research and drafting, but its benefits depend on how jobs are redesigned and whether employees receive the training and support to use it well. Poorly planned roll-outs can add checking work, confusion and concern about job security instead of easing workloads.

For companies, the message is clear: adoption figures alone are not a success measure. Businesses need goals, reliable data, staff training, safeguards and ways to track time saved, costs reduced and service improved. AI should solve defined problems, not be introduced simply because competitors are doing it.

The opportunity remains significant, but the next challenge is turning experimentation into responsible, well-managed use that delivers results people can actually see.

Version 1.0. Kobo Daily stories are edited before publication and updated as facts change. Spotted an error? Tell our editors.