Shalendra Singh and Chris Shortt agree that AI has two jobs: making a positive impact on the business, and people’s careers.
According to experienced retail CIO Chris Shortt, you have to work in aggregates when it comes to AI. That’s because the technology historically didn’t exist to work at a granular level, meaning that you need to aggregate wherever you can. In addition, he said, “people cannot deal with too much all at once. They need aggregation of information and data to try to get to some sort of decision and way forward.”
On AI in retail, Chris added that AI has been very effective in compiling the underlying datasets at a much larger scale, which has informed decision-making in a much more reliable and less biased way than aggregates did.
Bringing people along
In the African context, Chris explained that the retail space’s biggest goal is to be a net employer. AI moves us into a place where we start replacing people, when we should rather be using the technology to augment them, and help them be much better at what they do, he explained. “It’s all about nudging people to do something because you spotted something they may not have seen.
“The use of AI in decision-making is around formulating a proposition (the unique value that AI can bring to your business), and introducing that value into other use cases where you are helping people in the process of executing tasks, and becoming better at them. Particularly in the supply chain, you can augment someone and help them make decisions without relying on dashboards, by just prompting them on a task or model and how to deal with that the best way.”
How you start to make decisions in tasks, plan better, and maximise your assets are some of the considerations you need to make, said Chris. In retail, he says, inventory and merchandising are often your biggest costs, and a lot of time goes into putting systems in place that help optimise the use of that asset and manage costs, and vice versa.
“There are ways of doing that in the various parts of the value chain, particularly in the planning space, working at a granular level, and using AI to point out and isolate the trends that you don’t see when you’re working at an aggregate bias method – that way you’ll start to execute tasks better.
“How will all of that play out? It’s too early to tell. Especially if you focus too much on the ROI of each AI tool, you will probably never get off the ground with anything. It’s about weighing the two: identifying the value that comes out where you did apply these technologies versus what would have happened if you didn’t, because it only required the task to be executed properly in the first place.”
The challenges that come with AI
According to Shalendra, the biggest obstacle with AI is managing people’s perceptions about it and fears around it taking their jobs. “Ideally, AI should add value, and to a certain extent be a partnership between the technology and the people,” he says, “adding value to the business as a whole, while maintaining the human aspect, and people embracing it rather than fearing it.”
A manager constantly having the thought of AI replacing their jobs to some extent is even bad for the manager’s impact or interactions with customers. “It’s important that people see AI as something that can make them better at their jobs, getting better sales and their store running better. [There should be] a symbiotic relationship between your staff and these various AI and machine learning models.”
Chris said he agreed with the sentiments of UCT’s Kosta Kontos on building AI teams and shortage of skills in this regard, saying that the availability of skills is already slim to begin with thanks to global competition for those skills, and the need to develop that skill in a low-cost economy such as ours. There is high demand for those skills and people who have them are being snapped up very quickly.
“Building and retaining skill in an area that is moving very quickly is a significant challenge. There are two important skills needed: the people who can make the technology usable and workable, but another dimension is the education and awareness of people who might be the consumers of whatever is built.
“It’s the age-old challenge of trying to get people to trust system-generated outcomes that are statistically driven and scientifically founded in a space where people don’t have that mathematical understanding. And in the areas of planning and strategising, you need to create an appetite and the opportunity that exists, and tap into the people within the organisation who can spot that opportunity.”
Ethics and AI
According to Shalendra, IT as a whole is one of the most ethical professions in the world and that is because of data. For example, everyone in the room had access to their HR system and payroll system, therefore IT is generally perceived as an ethical discipline.
“The people running it are who you have to trust. That’s where the ethics come in,” he said.
“Our biggest focus at the moment is creating jobs, and the ethics piece comes in terms of AI not replacing them, but using that technology to expand the business. So our goal at the moment is to go to 40,000 employees, over the next 10 years, make a positive contribution to the country, and ensure that technology doesn’t negatively impact people’s lives,” Shalendra concluded.
