Executive Briefings

Opinion: In the Counterintuitive World of Supply Chain, AI Is a Force for Good

Planners working with complex supply chains will find that artificial intelligence systems will free them for more interesting work — and could even lead to rising in the company ranks.

In the Counterintuitive World of Supply Chain, AI Is a Force for Good

Society has mixed reactions about the forward march of artificial intelligence (AI). People naturally and often reasonably fear AI-based systems when they directly replace human tasks, say, on a factory assembly line or behind the wheel of a truck. There’s no denying that in the short term AI will affect some people’s jobs and require them to learn new skills.

Enthusiasm for AI is more prevalent, however, in situations where it can step in to solve challenging problems where counter-intuitiveness and complexity eludes humans and traditional systems. This is the case with supply chain planning. And we’ve also discovered that applying AI systems in complex supply chains improves planners’ working lives by freeing up their time for more interesting “human work,” raising the visibility of their contributions to company goals and even helping them earn promotions.

Why planning became so complex

Several relatively recent forces have made manual supply chain planning nearly impossible. On the demand side, the Internet, Amazon and mobile computing conspired to give consumers unprecedented information, access and channels to find, compare and buy products from virtually anywhere in the world. On the supply side, manufacturers today source labor, raw materials and components from more countries than ever.

These forces pose myriad challenges. The change in demand patterns puts companies under pressure to offer greater product choice and adjust their processes and policies to meet short, frequent and erratic demand cycles. With so many variables to consider on the supply side — fuel costs, exchange rates, lead times, storage costs, ethical sourcing, and so on — it becomes impossible for planners to consistently make optimal either/or trade-off decisions. A common example is the decision to either produce fewer items or run a promotion to sell more inventory. Another might be whether to order parts from a local supplier or save on unit cost by ordering from China but accept long lead times and high transportation costs. People’s best efforts to optimize these complex supply chains manually or using older systems end up being extremely time-consuming and frustrating.

Complex yes, but also an opportunity

In chaos however, some always find the opportunity to gain an advantage. Supply chain guru Martin Christopher, professor of marketing and logistics at the UK’s Cranfield School of Management, first said back in 2002, “Supply chains compete, not companies.” However, until recently there weren’t too many ways to creatively build a competitive supply chain. Companies were usually faced with costly trade-offs like: either limit product choice and service levels or throw large sums of cash at waste and inventory write-offs. Even mighty Amazon took the second approach (and arguably still does to an extent) in order to build and maintain market share in certain areas.

Fortunately, thanks to automation, companies no longer have to make potentially fatal choices like these. Technologies have been evolving to the point that they are affordable and accessible for companies of all sizes seeking to build truly responsive, competitive supply chains that are also lean and efficient. AI boosts the accuracy and predictive capabilities of these systems by tapping into a wide range of data sources not traditionally factored into supply chain planning, like CRM data, social feeds (which give valuable clues about sentiment) and even weather forecasts.

Why AI is right for supply chains

How is that AI can help supply chains yield high service levels and operate leanly? It’s because the process of optimizing a supply chain, while counter-intuitive to people, is ideally suited to algorithms.

Any planner who has ever learned supply chain theory by playing the “beer game” understands this well. In this classic simulation game, groups of people are given control over different parts of a supply chain, each of which can affect the overall outcome through their order quantities. Each group is further influenced by decisions made by the other teams. The lack of coordination and the incorrect decisions people make in response to other teams’ actions invariably results in the “Bullwhip Effect” (inventory shortfalls and overstocks across the supply chain). Even after planners learn how the Bullwhip Effect works in theory, they often fail to manage it in practice.

Another way to understand the Bullwhip Effect is to consider the effect of oversteering when driving a racing car or playing a “Formula One” video game. An experienced driver knows that after they turn the wheel, even if nothing happens, they need to stop turning to give the car a chance to respond. The novice’s instinct, however, is to keep turning the wheel until they see a response. This oversteering action causes the car to spin out, crash or skid off the track. If everyone on the racetrack oversteers, chaos and multiple crashes ensue. This is what happens in supply chains: different groups of people “oversteer” in response to the actions of the others causing inventory to pile up in some parts of the chain and fall short in others.

The Bullwhip Effect is amplified when people are motivated and rewarded by another human invention — “siloed” key performance indicators (KPIs). This happens, for example, when procurement people are rewarded for negotiating a low unit price on a raw material item through a bulk discount. This might be favorable when looking at the single order in isolation, but if excess items go unused, the inventory will eventually need to be disposed of and written off.

Automated planning systems enhanced by AI overcome these problems. They look dispassionately at supply chains as systems, making trade-offs that optimize against overall business objectives, like, for example, achieving a 99 percent service level or a 20 percent profit margin.

Improving business outcomes…

In every situation where our customers enhanced their supply chain planning using AI, we have measured impressive outcomes. In one example, the Italian dairy producer Danone achieved a 20 percent reduction in forecast error, a 30 percent reduction in lost sales and grew service levels to 98.7 percent. In another, the $7bn luxury eyeglass company Luxottica Group reduced manual planning by 50 percent, cut stock levels by 10 percent — all while maintaining the same high service levels.

…and personal ones

While we expected smarter systems to yield business improvements, what’s been really surprising is how much they also benefited the people working in supply chain teams. When planners no longer need to manually crunch numbers and double-check calculations, they are free to carry out the more “human” aspects of their jobs like gaining and applying business and market knowledge and taking more prominent and strategic roles in cross-functional, collaborative business initiatives like sales and operations planning (S&OP).

Andrew Lewis, head of global supply chain for the electronics supplier RS Components, summed up in a video the benefits his team gained by separating the machine’s job (creating the statistical forecast) from the planners’ job (enriching the forecast). He explained: “I want my demand planners to be ‘people people.’ I want interpersonal skills. I want people who can bring the sort of things that the system can’t probably know, and to do that I need to have a decent statistical forecast in the first place. Because I know it’s going to be massively improved by interaction with people.”

In a couple of our customer scenarios, junior planning team members have been elevated into more interesting, senior roles. Shamir Optical, a global eyewear manufacturer, promoted a junior data entry administrator — who took the initiative to master, tailor and creatively apply our software over several years — to head of supply chain. The UK’s most popular coffee retailer, Costa Express, redeployed a team of harried planners, whose main functions were number crunching and “logistics” (often driving emergency coffee supplies out to their partner sites). Those same people became Brand Excellence Advisors, working more proactively to help Costa Express’s partners grow sales, run more effective promotions and provide better service.

A force for good in supply chain

If intelligent machines don’t improve the lives and prosperity of the human race, there is little point in them. As responsible business leaders, it is incumbent on all of us to manage the introduction and use of AI technology in such a way that benefits employees and wider society. When AI replaces jobs, there must be concerted efforts between government and industry to support and re-skill people.

In supply chain planning, where humans have historically fought an uphill battle, AI technology is proving to be a force for good. We’ve seen it has not only consistently improved business outcomes, but the quality of people’s working lives.

Resource link: ToolsGroup

Society has mixed reactions about the forward march of artificial intelligence (AI). People naturally and often reasonably fear AI-based systems when they directly replace human tasks, say, on a factory assembly line or behind the wheel of a truck. There’s no denying that in the short term AI will affect some people’s jobs and require them to learn new skills.

Enthusiasm for AI is more prevalent, however, in situations where it can step in to solve challenging problems where counter-intuitiveness and complexity eludes humans and traditional systems. This is the case with supply chain planning. And we’ve also discovered that applying AI systems in complex supply chains improves planners’ working lives by freeing up their time for more interesting “human work,” raising the visibility of their contributions to company goals and even helping them earn promotions.

Why planning became so complex

Several relatively recent forces have made manual supply chain planning nearly impossible. On the demand side, the Internet, Amazon and mobile computing conspired to give consumers unprecedented information, access and channels to find, compare and buy products from virtually anywhere in the world. On the supply side, manufacturers today source labor, raw materials and components from more countries than ever.

These forces pose myriad challenges. The change in demand patterns puts companies under pressure to offer greater product choice and adjust their processes and policies to meet short, frequent and erratic demand cycles. With so many variables to consider on the supply side — fuel costs, exchange rates, lead times, storage costs, ethical sourcing, and so on — it becomes impossible for planners to consistently make optimal either/or trade-off decisions. A common example is the decision to either produce fewer items or run a promotion to sell more inventory. Another might be whether to order parts from a local supplier or save on unit cost by ordering from China but accept long lead times and high transportation costs. People’s best efforts to optimize these complex supply chains manually or using older systems end up being extremely time-consuming and frustrating.

Complex yes, but also an opportunity

In chaos however, some always find the opportunity to gain an advantage. Supply chain guru Martin Christopher, professor of marketing and logistics at the UK’s Cranfield School of Management, first said back in 2002, “Supply chains compete, not companies.” However, until recently there weren’t too many ways to creatively build a competitive supply chain. Companies were usually faced with costly trade-offs like: either limit product choice and service levels or throw large sums of cash at waste and inventory write-offs. Even mighty Amazon took the second approach (and arguably still does to an extent) in order to build and maintain market share in certain areas.

Fortunately, thanks to automation, companies no longer have to make potentially fatal choices like these. Technologies have been evolving to the point that they are affordable and accessible for companies of all sizes seeking to build truly responsive, competitive supply chains that are also lean and efficient. AI boosts the accuracy and predictive capabilities of these systems by tapping into a wide range of data sources not traditionally factored into supply chain planning, like CRM data, social feeds (which give valuable clues about sentiment) and even weather forecasts.

Why AI is right for supply chains

How is that AI can help supply chains yield high service levels and operate leanly? It’s because the process of optimizing a supply chain, while counter-intuitive to people, is ideally suited to algorithms.

Any planner who has ever learned supply chain theory by playing the “beer game” understands this well. In this classic simulation game, groups of people are given control over different parts of a supply chain, each of which can affect the overall outcome through their order quantities. Each group is further influenced by decisions made by the other teams. The lack of coordination and the incorrect decisions people make in response to other teams’ actions invariably results in the “Bullwhip Effect” (inventory shortfalls and overstocks across the supply chain). Even after planners learn how the Bullwhip Effect works in theory, they often fail to manage it in practice.

Another way to understand the Bullwhip Effect is to consider the effect of oversteering when driving a racing car or playing a “Formula One” video game. An experienced driver knows that after they turn the wheel, even if nothing happens, they need to stop turning to give the car a chance to respond. The novice’s instinct, however, is to keep turning the wheel until they see a response. This oversteering action causes the car to spin out, crash or skid off the track. If everyone on the racetrack oversteers, chaos and multiple crashes ensue. This is what happens in supply chains: different groups of people “oversteer” in response to the actions of the others causing inventory to pile up in some parts of the chain and fall short in others.

The Bullwhip Effect is amplified when people are motivated and rewarded by another human invention — “siloed” key performance indicators (KPIs). This happens, for example, when procurement people are rewarded for negotiating a low unit price on a raw material item through a bulk discount. This might be favorable when looking at the single order in isolation, but if excess items go unused, the inventory will eventually need to be disposed of and written off.

Automated planning systems enhanced by AI overcome these problems. They look dispassionately at supply chains as systems, making trade-offs that optimize against overall business objectives, like, for example, achieving a 99 percent service level or a 20 percent profit margin.

Improving business outcomes…

In every situation where our customers enhanced their supply chain planning using AI, we have measured impressive outcomes. In one example, the Italian dairy producer Danone achieved a 20 percent reduction in forecast error, a 30 percent reduction in lost sales and grew service levels to 98.7 percent. In another, the $7bn luxury eyeglass company Luxottica Group reduced manual planning by 50 percent, cut stock levels by 10 percent — all while maintaining the same high service levels.

…and personal ones

While we expected smarter systems to yield business improvements, what’s been really surprising is how much they also benefited the people working in supply chain teams. When planners no longer need to manually crunch numbers and double-check calculations, they are free to carry out the more “human” aspects of their jobs like gaining and applying business and market knowledge and taking more prominent and strategic roles in cross-functional, collaborative business initiatives like sales and operations planning (S&OP).

Andrew Lewis, head of global supply chain for the electronics supplier RS Components, summed up in a video the benefits his team gained by separating the machine’s job (creating the statistical forecast) from the planners’ job (enriching the forecast). He explained: “I want my demand planners to be ‘people people.’ I want interpersonal skills. I want people who can bring the sort of things that the system can’t probably know, and to do that I need to have a decent statistical forecast in the first place. Because I know it’s going to be massively improved by interaction with people.”

In a couple of our customer scenarios, junior planning team members have been elevated into more interesting, senior roles. Shamir Optical, a global eyewear manufacturer, promoted a junior data entry administrator — who took the initiative to master, tailor and creatively apply our software over several years — to head of supply chain. The UK’s most popular coffee retailer, Costa Express, redeployed a team of harried planners, whose main functions were number crunching and “logistics” (often driving emergency coffee supplies out to their partner sites). Those same people became Brand Excellence Advisors, working more proactively to help Costa Express’s partners grow sales, run more effective promotions and provide better service.

A force for good in supply chain

If intelligent machines don’t improve the lives and prosperity of the human race, there is little point in them. As responsible business leaders, it is incumbent on all of us to manage the introduction and use of AI technology in such a way that benefits employees and wider society. When AI replaces jobs, there must be concerted efforts between government and industry to support and re-skill people.

In supply chain planning, where humans have historically fought an uphill battle, AI technology is proving to be a force for good. We’ve seen it has not only consistently improved business outcomes, but the quality of people’s working lives.

Resource link: ToolsGroup

In the Counterintuitive World of Supply Chain, AI Is a Force for Good