The barrier to AI-led digital transformation was never the technology

By Krishna Chetty, Executive Managed Services and Cloud, Altron Digital Business

LONDON (Capital Markets in Africa) – The biggest barrier to digital transformation is rarely the technology. It’s rather the distance between the people who decide what to build, which systems to install, and which processes to implement and the people who’ll actually use it. 

It’s when that distance is closed that most of what is labelled “resistance to change” is no longer seen as a barrier to digital transformation because it never was – it was simply a disconnect between what was needed and what was determined was needed. 

This isn’t a failure of intent but is rather a structural problem. The further a decision sits from the implementation layer of a company, the easier it becomes to mistake a clever solution for a good one – and the less likely anyone is to notice the wrong thing has been built until it’s already live. 

The discipline that matters here is simple, if unfashionable. It’s to spend real time with the people doing the work, and they will tell you what’s broken without needing to be asked twice. Not a survey or a workshop; it’s proximity that counts. 

Every transformation that has actually stuck follows the same pattern: the fix wasn’t designed in isolation and handed down. It was shaped by the person who was going to live with it. 

This maxim holds true regardless of where in the company it should be employed; from spending time in a call centre through to discussing an accountant’s concerns about the outputs of the software they are using. And it’s about listening to your people, because friction left unresolved is what quietly limits both productivity and profit. 

The AI lesson 
Generative AI made this visible faster than almost anything before it. Long before formal policy caught up, employees across industries were already using these tools to draft faster, analyse data, and work through tasks that used to take considerably longer. 

The instinct from a distance is to treat that as a risk to shut down – and there is real risk in it because sensitive data moving through ungoverned tools is a genuine problem, not a hypothetical one. 

But shutting it down misses what the behaviour is actually saying. People weren’t being careless; they were solving their own friction faster than the organisation could supply an approved answer. That isn’t defiance. It’s a signal and is one of the clearest examples of a human barrier dissolving itself before leadership had even noticed it was there. 

The organisations getting this right aren’t the ones banning the behaviour or ignoring it. They’re the ones building the approved version around what people had already proven they need, matching the tool to the guardrails so that employees work within a controlled environment rather than applying an outright ban. 

Staff will adopt any tool they deem necessary to do their jobs; this is a universal truth. As a result, what needs to shift is that the risk moves from unmanaged to managed, without anyone losing the technological tool that was making their day easier. 

In fact, our recent Employee Technology Experience Index found that nearly a third of financial services employees are already sourcing their own AI tools, showing that a shift to employees using tools outside of the company’s controlled environment isn’t hypothetical – it’s already happening. 

Ubiquitous experiences 
This isn’t unique to knowledge work. Frontline operations, a call centre fielding thousands of alerts a month, most of them false, show the identical dynamic. The team closest to the noise usually knows exactly which signals matter long before that knowledge reaches a dashboard. 

The fix is rarely a new platform. It’s giving that frontline judgement somewhere to go and rather redeploying effort toward what’s actually real, rather than adding more process to filter what already shouldn’t need filtering.  

None of this requires a fresh transformation programme each time. It requires choosing one process someone deals with daily, asking what would make it easier, and following through. A smooth-running environment isn’t built by adding oversight from above. It’s built by removing what’s actually in the way and that’s rarely visible from an isolated management position. 

Being human 
If there’s one barrier standing between most organisations and transformation that actually sticks, it’s that leadership sits structurally furthest from the people doing the work, and the fixes born of that distance are usually the ones nobody asked for. 

The way through isn’t more governance from the top. It’s staying close enough to notice what people are already solving for themselves and having the humility to build governance into the process and around that, rather than over it. 

Technology will keep moving faster than most organisations can govern it. What won’t change is that transformation only sticks when the people closest to the work are treated as the source of the answer, not the last people consulted about it.