Something is happening in the South African market that doesn’t always surface in the enthusiasm around AI. Customers are investing. Pilots are running. Licences are in place. And twelve months later, the question is why it hasn’t delivered. It is not the technology.
What we’re seeing across the market is a large number of organisations that have completed a pilot and assumed the work is done. They’ve watched the inbox searches and the meeting summaries. Some have identified a specific use case that briefly excited them. But they haven’t done the harder work that actually makes AI useful, and the investment continues to underperform as a result.
That harder work is not glamorous. It starts with a clear question: what is the business trying to achieve, and how does that translate into something measurable? From there it extends into data readiness, governance, and the cultural question of whether people will use the tools they’ve been given, not just acknowledge them. Most organisations haven’t made much progress on any of it. They are running fragmented pilots inside individual departments, without a coherent view of how those pilots connect to business outcomes. The spend has been real. The returns, often, have not.
I say this as someone who has lived the same journey. My first instinct when AI became unavoidable as a conversation was scepticism. I thought it was paper fire. I ran the free version for months before I understood enough to invest in the licence, and even now I know I’m not using it optimally. What has slowed our own broader rollout isn’t access to tools. It’s governance. We want to get the controls right before we scale, because rolling out broadly, losing confidence, and having to walk it back costs far more than moving carefully.
A small law firm we spoke to recently hadn’t thought about any of this in advance. They had one problem: they wanted to produce professional-looking material to compete with firms that had marketing departments they couldn’t afford. They started there. It worked. And it gave them a foundation to build on. That is the right sequence: a specific problem, a clear outcome, then a question about what comes next. Not a pilot that ends with a shrug and a renewal decision.
There’s also a competitive dimension that’s easy to underestimate. Not adopting AI properly is not a neutral position. Businesses that have got their implementation right are already operating faster, delivering more, and getting quotes out the door in less time. Those that haven’t aren’t just missing an opportunity. They’re accumulating a disadvantage that will be harder to close the longer it remains.
And there is something larger at stake for African economies than any individual business outcome. I am convinced that AI is a genuine step change, not just another technology wave. We’ve seen before how African markets can move past infrastructure constraints and adopt what’s current rather than what’s conventional. That opportunity is real. But it’s only accessible through the work most organisations are still treating as optional.
The pilot was the easy part. What comes after it is the question worth answering now..
- Senzo Mbhele, managing director, Cloud On Demand
