Why trust is key for artificial intelligence adoption in the consumer goods supply chain
Despite a promising future, adoption of artificial intelligence (AI) in consumer goods manufacturing and supply chain management has been much slower than in the technology, retailing and financial services sectors due to a lack of data for analytic tools to work on, according to a supply chain expert.
“AI is about the collection and analysis of data and the application of insights gained. So far the most successful applications of AI are in facial and voice recognition, cartoon animation, medical diagnostics and automatic navigation,” Hau Lee, chairman of the board of Fung Academy, said in an interview.
The academy is a business unit under Fung Group focusing on staff training on technology adoption, as well as fostering innovation and new technology applications across the group’s businesses.
Lee is also a professor of operations, information and technology at the Stanford Graduate School of Business, and had co-founded several supply chain and price optimisation software firms in the United States.
Lee noted that the application of AI in supply chain management has been slower than in other industries because industry participants are reluctant to share their operating data.
“What’s more, each participant only has access to data limited by the products they make, the geographical markets they serve, in a confined section of a long supply chain,” he said.
Even so, Lee said data analytics has big potential in supply chain management since “even a 1 per cent cost saving in business-to-business dealings can amount of hundreds of millions of dollars”.
Even in the broader business world, adoption of AI is only at an experimental stage, according to a McKinsey survey of 3,000 executives in 10 nations across 14 industries.
“Many firms say they are uncertain of the business case or return on investment,” said the report penned by partners of the global management consultancy. “A review of more than 160 cases shows that AI was deployed commercially in only 12 per cent of the cases.”
Still, it said AI “is poised to unleash the next wave of digital disruption” and companies should prepare for it, since early adopters have achieved higher profit margins and the gap with non-adopters is expected to widen.
So far, most of the investment in AI is spent by global technology firms.
McKinsey estimated that tech giants such as Google and Baidu spent some US$20 billion to US$30 billion on AI last year, of which 90 per cent was on research and development and the rest on acquisitions of intellectual properties or companies.
Machine learning – which enables computers to learn and adapt through experience without explicit programming – accounts for the largest share of the investment.
To make data sharing work in the long supply chain of manufacturing, Lee said it was key to have a platform that allows all participants to share their data in a secured manner for mutual benefit.
With a global network of 15,000 factories and 8,000 retail and brand owners as customers, Lee said Fung Group, having built trust over many years with its business partners, could act as a data aggregator and AI insights disseminator for customers and suppliers.
“Our dream is to help our customers come up with a winning collection of products based on the latest market intelligence, and be able to work with the factories to offer our customers discounts if they are willing to use excess materials stock from last season’s production, for example,” he said.
“This requires the group to demonstrate that all data providers will reap benefits and that the data will not be leaked to unintended parties”.
However, he admits this “dream” of an integrated supply chain data sharing platform is a huge challenge and will take a long time to achieve, even for an established player with over 110 years of operating history.
Still, Lee said even without the sharing of internal data such as retailer inventory and available factory production capacity, plenty of data is already available to Fung Group’s supply chain management unit Li & Fung that can be used to add value and save costs.
“For example, Li & Fung’s database allows us to allocate orders to the most suitable and efficient factories for the manufacturing of different products, and give practical advice to factory workers on efficiency improvements,” he said.