AI and the Supply-Chain

Artificial Intelligence is radicalising the modern supply-chain.Here’s how

The strategic centrality of artificial intelligence applications has been much apparent over the past few years – especially recently, in its role in helping the economy phase itself out of some of the most complex issues brought about by the global coronavirus pandemic. Of these issues, few will show as stark a contribution of AI ingenuity than recovering the supply-chains pushed to their extreme limits owing to market volatilities exacerbated by the pandemic. Consider these examples from the MIT Technology Review:

“Google is developing supply-chain digital twins that the car maker Renault announced it had started using in September. International shipping giants like FedEx and DHL are building their own simulation software. And AI firms like Pathmind are creating bespoke tools for anyone who can pay for them.”

It’s not just the global giants of Amazon or Walmart that have benefitted, however, artificial intelligence technologies have been unequivocal in strengthening our global supply-chains in more ways than one. This holds especially true today, considering the onus being placed on reducing their environmental impacts as well – something that has triggered firms worldwide to concentrate on supply-chain resilience and the sustainable optimisation of flows. McKinsey thus writes:

“An integrated end-to-end approach (based on AI) can address the opportunities and constraints of all business functions, from procurement to sales. AI’s ability to analyze huge volumes of data, understand relationships, provide visibility into operations, and support better decision making makes AI a potential game changer. Getting the most out of these solutions is not simply a matter of technology, however; companies must take organizational steps to capture the full value from AI.”

The Data-driven supply-chain

Some of the AI-based solutions in vogue today to help companies achieve improved performance in supply-chain management include the use of demand-forecasting models, dynamic planning optimisation, correlation analysis, integrated business planning, prediction models and the automation of flows – all of which have so far allowed organisations to cut logistics costs by almost 15%, improve inventory levels by 35% and service levels by almost 65% compared to slower-moving counterparts.

• Demand Analytics and Network Planning: Predictive analytics forecasting future demand levels based on current and past sales’ data can help at various sale points including retailers, stores or distributors, even taking into consideration aspects such as holidays, weather forecasts or integration with promotional events. Additionally, it is imperative to ensure that the inventory and manufacturing facilities are properly networked with smooth flow paths to ensure fulfilling customer demands at the lowest costs.

US-based Symbotic, for example, “designs, builds and tests AI-powered robots that provide flexible manual or fully automated solutions based on a company’s products, operational flow and customer needs”; while Havi, another AI firm based out of the US, designs predictive analytics models for supply-chain management (including planning, sourcing, optimisation and data management) and logistics, (including warehousing, freight management, procurement and distribution.)

• Finished Inventory Optimization and Replenishment Planning: This predictive analysis gives a clear estimate of how much inventory there should be and how it ought to be positioned to make the budgeting process more optimised; and how much safety stock needs to be maintained to combat the effects of fluctuating demand. Replenishment planning then gives a clear idea of shipments, thereby allowing seamless integration between channel, retailer and distributor.
• Procurement and Transportation Analytics: One of the primary aspects of the supply-chain process, procurement analysis allows the accrual of raw materials from high-quality suppliers, taking into consideration aspects like scoring models, vendor quality as well as overall long-term stability. Transportation analysts then visualise the best possible route to move products down the supply-chain, including aspects such as shipping and backhaul routes, scheduling techniques and constraints and compliances of the above.

US-based Echo Global Logistics and Uptakeare among firms that have made great strides in this regard.  Uptake, for example, “uses AI and machine learning to analyze data for telematics with the goal of predicting failure in order to reduce downtime for a variety of vehicles and machinery, including trucks, cars, railcars, combines and planes.”

Possible Pitfalls

The transition to an AI-based supply-chain, however, is not the smoothest one, and firms need to be aware of how to navigate several kinds of implementation challenges.

• Identifying the Value Strategy: As a first step, firms need to identify and prioritise all pockets of value creation across functions including manufacturing, procurement and logistics. Research has found, however, that only about one-third of the firms surveyed perform independent diagnostics right at the outset.

“Clearly defining a digital supply-chain strategy helps support the company’s business strategy and ensures better alignment with its digital program. In addition, a solution-agnostic assessment enables companies to identify the process redesign, organizational changes, and capabilities required to boost performance as well as create a strategic road map”, according to McKinsey.

• Solution Design and Vendor Selection: Secondly, given the complexity of the various nodes of a supply-chain, it is often not possible to narrow down options to a single provider to meet all needs; each vendor usually sticks to a specific end-to-end goal.

In such a case, McKinsey opines, “solution design and vendor selection can help support the digital supply-chain strategy. Often, the best approach is a combination of different solutions from different providers, implemented by different systems integrators.”

• Implementation and Systems Integration: A major challenge that firms must overcome is to implement these solutions within budget and to scale, while properly addressing value-creation as the primary goal. Only about one-fourth of global supply-chain leaders reported that they felt their objectives were properly aligned with their systems integrators.
• Change Management: Even whilst making the digital transformation to an AI-powered supply-chain, firms must be aware of aspects such as change management, capacity building and organisational structure. Only about 13% of the surveyed executives believe that their firms are prepared to address the issue of skill gaps. McKinsey opines:

“To ensure adoption of new solutions, companies must invest in change management and capability building. Employees will need to embrace new ways of working, and a coordinated effort is required to educate the workforce on why changes are necessary, as are incentives to reinforce the desired behaviors.”

The benefits of applying data science to supply-chains are rather clear. It guarantees improvements to accuracy, management and pattern recognition, whilst optimising costs and performance. A well-managed production line is the backbone to expansion, allowing for newer products and as well as an enhanced supply-chain with smoother workflow, reduced latency and balanced compliances and constraints.

Whilst remaining wary of the possible pitfalls is crucial, supply-chain management is an application of data science that is all set to boom now and well into the next decade, especially with global trade and commerce gradually opening up post the coronavirus pandemic.

AI and the Supply-Chain