Adrian Adams, from the Council for Scientific and Industrial Research (CSIR), contributed to this thought leadership piece
The South African Procurement and Supply Chain (P&SC) community is yet to grasp the application of Artificial Intelligence (AI) and Machine Learning (ML) technologies in driving business value. According to Letsema’s 2018 P&SC Research Report, 65% of executives interviewed indicated little or no understanding of these important emerging technologies.
Adrian Adams, of the CSIR, says Africa has the potential to make a significant contribution to global AI & ML revolution by developing unique solutions to problems impacting society and business on the continent and globally. Like electricity, the AI & ML revolution is predicted to transform the world we live in, creating opportunities everywhere: from manufacturing, supply chain, finance, and healthcare to education, agriculture, and beyond.
In the context of growing African economies and building competitive, sustainable businesses this sparks a key question that executives are grappling with:
“What role can ML & AI play in optimizing Procurement and Supply Chain in the South African and African context?”
Unpacking the answer starts with understanding the buzzwords :
- AI systems are considered to be machines or software systems that can perform human level intelligence tasks and learn from patterns in data or experience
- Machine learning is considered to be a sub-field of AI, focused on the development of machines that learn from data on their own without being explicitly programmed
Companies today are operating in complex markets, with new operating models and partner ecosystems generating vast sources of seemingly disparate data. Traditional quantitative and analytical techniques may start to show gaps, while insights or risks may be harder to assess and plan for. In Africa, powerful AI/ML techniques can be applied to draw patterns and analyse alternative solutions to address these gaps, compared to more traditional techniques.
Contrary to some popular literature and thinking in the market, we believe that AI & ML techniques can be viewed as evolutionary, and not disruptive, in the sense that AI/ML techniques have already been applied for a number of decades in business. Examples include hand writing recognition and optical character recognition in postal systems and banking, courier companies optimising delivery of parcels, automation, optimisation of aircraft routing, oil exploration and optimisation of chemical factory processes. These examples of innovation have not been touted as a ‘big bang’ moment where AI/ML is ready to transform all industries.
While it is a positive step for P&SC executives and teams to experiment with and deepen knowledge across these emerging technologies, technology for technology’s sake is not a good business strategy. The application of AI & ML needs to be applied with the intent to solve specific business challenges, based on available information.
To bring these points together in a holistic way, the framework below is useful to articulate what is needed for the application of AI & ML in P&SC.
To unlock the true value of AI & ML, these three ingredients need to be incorporated simultaneously:
- Key Business Questions: Clear articulation of a business challenge or opportunity being solved for. Typically involves companies wanting to apply data-driven decision making related to improving business performance measures (including cost, risk, growth)
- Company and External Data:
- Company data: Internally generated business data gathered from operations
- External data: Includes macro-economic factors that influence a company’s performance, such as commodity prices, forex, consumer price index, interest rates, weather patterns, industry data, consumer behaviour indicators, etc.
- AI & ML algorithms: Powerful techniques exist and new advances in AI/ML continue to be made that create new opportunities to apply and solve key business problems.
If we were to look for an example of how this is achieved, Lenovo offers an informative case study below,
Leveraging the Framework: A Case in Point(1)
Lenovo is a multinational technology company, with four devices sold per second. Lenovo Brazil embarked upon a mission to predict sell-out volumes – their key question – by using ML.
Predicting sell-out volume is not a new problem. The issue is the volume of data that requires processing. A total of 59 variables were identified, comprising of both internal and external data, such as product and competitor pricing, historic sell-out periods, Google Analytics rankings, etc.
By processing this data using various ML algorithms hosted on Amazon Web Services (AWS), Lenovo were able to arrive at a solution where sell-out volume predictions are now used to define action planning within the company. An accuracy rate of 90% is now achieved in sell-out volume predictions for up to four weeks in advance. This is an example of how ML can be used when all the key ingredients are present.
While the clock is ticking…
According to the Letsema’s 2018 P&SC Research Report, the benefits of AI & ML are not yet fully comprehended by South African SC executives.
The evolving AI & ML landscape should not deter P&SC executives from challenging themselves and their teams from starting to understand the opportunities and practicalities of leveraging these important technologies. Relatively small steps can be initiated today, so that companies are not left with only mountains to climb.
Consider the following:
- AI & ML solutions need to be trained from historic data. Even if a transformation approach or strategy has not been fully formed, start collecting data, keep it clean through proper data governance and make it accessible to the organisation (e.g. on cloud platforms)
- Critically and creatively explore the potential opportunities for AI & ML in P&SC – across industries – and the potential for disruption. Perhaps initiate a strategic review process
- Develop and align on an AI & ML Strategic Roadmap for P&SC
- Build capacity and pilot identified AI & ML opportunities. Start small to prove the concept, learn lessons and build trust, and scale
1Source: AWS (n.d.) Lenovo Brazil Improves Supply and Demand Balance with DataRobot, AWS. Available at: https://aws.amazon.com/partners/apn-journal/all/lenovo-datarobot/