Kathy Gibson is at IDC CIO Summit in Zimbali – Artificial intelligence (AI) is undoubtedly the next era of interest when it comes to future trends and it has a massive role to play in digital transformation.

“At IBM, we pay a lot of attention to digitalisation and using AI across the stack,” says Hein Badenhorst, client technology adviser at IBM

“We focus on AI capabilities as part of our day-to-day operations,” he adds. “There is a lot of manual intervention that we cannot afford.”

One of the areas that must receive focus in AI is the operating models, Badenhorst adds. “The traditional operating model is a range of processes run by people and supported by technology.

“But, today, the processes are run by technology and supported by people.”

However, very little attention is paid to strategies around the operating models used for AI.

Companies often opt for AI because software tends to be cheaper than people. But Badenhorst points out it raises new and different expenses.

For instance, how do you ensure people have the new skills needed to work with AI? These other issues have to be addressed.

Data scientists, on the other hand, tend to be free spirits who create software. But companies need to be able to trust the applications they created.

There could be significant impact on people and the company if the software doesn’t behave as it should.

Badenhorst adds that scale is a problem in the AI space. “You need to scale to suit your own organisation and also other companies.”

Bias is another issue that AI is prone to, and this has to be managed, he adds. “You need to have a constant feedback loop without manual intervention.”

Legacy is another important consideration, and can never be ignored, Badenhorst says.

IBM recently moved all American Airlines customer-facing systems into the cloud, co-ordinated into one event.

For organisations looking to move to new, AI-enabled systems it shouldn’t be a big bang, though, says Badenhorst. “I wouldn’t recommend that you lose your old version of the world.”