South Africa’s data analytics companies have advised that the retail community should adopt artificial intelligence (AI) as a business tool or perish.
Dewald Lindeque, business development director at Moyo Business Advisory, a Centurion-based specialist business consultancy and big-data analytics house, said the limited introduction of AI in South Africa had yielded “stunning” results and closely followed the successes of both online and bricks and mortar retailers internationally.
“We have seen companies on the brink of disaster turn their businesses around and go from loss-making enterprises to becoming profitable thanks to the introduction of AI to improve all aspects of their operations.
“Ranging from automating warehouses to ensuring adequate stock levels at their shops and implementing chatbots to interact with clients to enhance the customer experience, AI holds the promise of revolutionising the retail sector,” Lindeque said.
“The hard-pressed retail sector has been savaged by the economic downturn and about the only ray of hope on the horizon is AI coupled with effective data analytics.
“In addition to throwing a lifeline to the retail industry, it is also going to open up brand new horizons for consumers who will have a substantially more pleasurable shopping experience – both online and in brick and mortar retail outlets,” Lindeque said.
Dr Duncan Coulter of the University of Johannesburg’s Computer Science Academy said in most cases, retail enterprises already had access to vast amounts of valuable data that could kick-start the implementation of AI systems.
“There is no doubt that AI is going to become an ever more valuable part of our lives in virtually every sector and retail is certainly no exception. Chatbots interacting with callers online through text and voice have already become a valuable part of our lives giving advice and taking orders in a wide variety of enterprises.
“It holds immense promise for things like autonomous drones that are already delivering urgent blood supplies in a number of African states and will shortly be used by retailer Amazon for the delivery of parcels.”
Coulter said so-called push technology where shoppers in malls would receive messages from shops in the area offering special deals tailored to the specific likes of individuals based on their AI profiles was another aspect of AI that held great promise.
“Privacy issues would obviously have to be taken into account and the ideal situation would be to incentivize shoppers to opt into a scheme where they provide detailed data about their likes and dislikes which could then be tailored by an AI system to offer them special discounts on items on their wish list.
“The same system could also be used to track their shopping patterns both online and in bricks and mortar shops to make recommendations for additional items that would match a garment or fashion item they had selected.”
Lindeque said AI and accompanying data analytics are already being used to target advertising to consumers via Google and social media to target their specific likes and dislikes.
“Rather than spending millions on glitzy television ads that attempted to target everybody but in fact only managed to influence a very small percentage of viewers, with AI it is possible to create promotional material that you know will reach exactly the right audience with great accuracy.
Technologies like smart mirrors which are in fact large computer monitors with built-in censors that are capable of overlaying clothing items that might match whatever garment a punter was trying on are a little further in the future.
“They are routinely being used in Japan and China to boost sales but the technology is expensive. In China the giant e-commerce etailer Alibaba has been subsidising physical shops to put up these screens in their fitting rooms. In the 13 shops across China where the systems are in use, they have been phenomenally successful in boosting sales.
“Through near-field dongles embedded in the clothing the AI system driving the display screen knows which garment is being tried on and can then suggest additional fashion items that match the customer’s profile and supply a visual representation of what the item would like on the customer.