Artificial Intelligence, Automation, Features, Logistics, Logistics software

Is AI the future of supply chain?

AI and supply chain

Peter Jones, Managing Director of Prological, discusses AI’s prospects for improving supply chains, how earlier technological predictions have panned out, and why AI might be less transformative than many of its current boosters think.

In the late 90s and early 2000s, there was a deafening buzz in the supply chain industry about RFID (Radio-Frequency Identification) technology. According to Peter Jones, Managing Director of Prological, every supply chain magazine was brimming with stories about the transformative potential of this technology.

“There was constant talk about its capabilities and how it could transform the industry,” Peter recalls. At that time, the most significant hurdle was the prohibitive cost of smart RFID tags – the ones that can update information on them – as they necessitated an interactive chip to fulfil their touted potential; ‘the cloud’ didn’t yet exist.

Peter points out how industry heavyweights worldwide invested heavily in this promising technology. “Materials handling companies around the world set up their RFID centres,” he says. However, the gap between the hype and the reality soon became evident.

Back then, Peter was a card-carrying RFID booster. He helped organise a project at the University of Wollongong that aimed to quantify the value proposition of incorporating RFID tags into the LG supply chain network in Australia. The project, backed by significant Australian Research Council (ARC) and private funding, focused on Sydney. The synopsis was that an RFID smart tag be added to a product at the production line’s start and traced throughout the LG supply chain all the way to the end consumer.

“The idea was that we could use that RFID tag at the point of sale all the way through to the customer,” he says. “The benefits were perceived as being significant all the way through the value chain, including warranty and other retailer and customer uses.”

Despite demonstrating about a $2.5 million commercial benefit to LG just in the operation of the Sydney warehouse, RFID technology never quite took off to the degree Peter might once have anticipated.

“We proved its theoretical case, but the operational obstacles and the reality of the world to get it to work were too profound for the market at the time.”

Blockchain, like RFID, evidenced a similar gap between expectations and reality. Despite successful ‘Ocean Protocol’ pilot projects (as just one example of application) by companies such as Roche, Unilever and Johnson and Johnson, blockchain’s adoption in supply chain management has been less transformative than initially projected.

Peter explains the principle behind blockchain. “Once you have a set of data, this data can flow through all required channels from beginning to end without further data entry or data transfers, but the file is added to and updated at every data interchange,” he says. “You can’t modify the data retrospectively, therefore you create one source of truth; and in this is its potential power. This is the theory.”

However, reality checks in when considering the broader applicability of such a system. Peter points out that while blockchain can work effectively in controlled environments – such as a single product, single source of supply, and single destination – it falls short in broader business operations: “Blockchain requires perfect data. No businesses have perfect data across their operations yet and herein lies the gap between the possible and achievable.”

ET TU, AI?

Notwithstanding prior disappointments with RFID and blockchain, the supply chain industry is ready once again for the Next Big Thing. But will AI, too, betray our great hopes?

When asked what people mean when they talk about AI and the supply chain, Peter acknowledges the nebulousness of the term. “Some people think of fully human-like robots (humanoids) with lateral and logical thinking skills; for others it is simply about refined data analysis with machine learning and predictive mechanisms,” he says. “I think – at least in terms of supply chain – we’re more likely to see help coming from AI in the latter rather than in the former.

“The fact is, the basics of supply chain remain the same: there’s a product at point A that needs to get to point B, and people are involved in that process,” Peter says. “Until we have fully automated warehouses and transport systems, we’re a long way from achieving full automation. In other words, we’re not even close to the kind of AI that could think for itself to the degree that humans become unnecessary in supply chain.”

But that doesn’t mean AI isn’t having an impact in discrete areas.

“The area where I believe AI will have a significant impact quickly is in forecasting and inventory management, because these tasks are heavily algorithm-driven,” Peter notes. He cites Amazon’s sophisticated use of big data in combination with AI for inventory management as an example of the potential for AI in this area. “Using AI, they know where those customers live, and they can get even very low-demand items located correctly – say, a pair of socks that only three people have ever bought – to the right distribution centre before those three customers even know they need a new pair.” Anywhere in supply chain that has a critical predictive element will be well served by AI, and in combination with big data, will support ‘stepped change’ supply chain execution.

The introduction of AI into the supply chain industry prompts questions about its potential impact on the workforce. Will it necessitate new types of roles? Will it attract a younger generation to the supply chain industry? Or will it simply make tasks easier for less skilled individuals?

Peter thinks that while AI might lead to the creation of new roles, it’s not necessarily going to reshape the workforce or make the industry more attractive to a younger demographic. Advances in automation and advanced Warehouse Management Systems are already doing this at a rapid pace and it is hard to see AI significantly adding to the pace generated by the current wave of development.

The fundamentals of supply chain remain the same, and so it will attract people with skills and mindsets that have always fit supply chain, Peter says. Changes to the workforce will occur to the degree new workers are required with the technical skills to manage new technology.   

“If we have warehouses full of AMRs, the supply chain team now needs engineers and programmers for these AMRs,” he says. “That didn’t used to exist. They are now part of the supply chain team.”

Reflecting on his own 27 year career, Peter says that – in spite of some promised tech revolutions not panning out – “there have been massive changes I couldn’t have imagined back in the mid-1990s.”

By way of example Peter cites the capabilities of contemporary warehouse management systems (WMS).

“Think about the ability of a WMS to automatically reposition product around your warehouse across the seasonality of your year, getting your WMS preparing space for inbound product that’s still eight weeks away without any human intervention; or interleaving of task allocation by the WMS, leading to reduction in ‘dead-running’ in the warehouse,” he says. “Advanced WMS will do 95 per cent of these tasks. This is quite amazing, when you reflect on where we were a few decades back.”

Another significant evolution is the integration and convergence of systems. There was a time when connecting two different computer systems was a daunting task. The ability of various systems to seamlessly ‘plug into’ each other has now removed that concern.

“The integration of technology platforms – like an ERP system and a WMS or forecasting system or transport management system – is now as easy as connecting an extension cord,” Peter says. “Well not quite, but today connectivity is not a systems constraint.”

The capabilities of track and trace technology have also exceeded expectations, with significant advances in the B2C sector. Now, B2B operations are following suit, trying to emulate the B2C approach by providing extensive, real-time tracking information.

“It takes a lot to understand the complexity of the task that goes behind fulfilling an order and getting it to your place on time, seeing how many things had to happen, how many people were involved, and how many exchanges occurred with the product in a journey. And yet, it all happens with 98 or 99 per cent first-time accuracy – and you have all of that information in your pocket.”

In the end, Peter thinks that it’s those technologies which consistently improve on supply chain fundamentals – rather than trying to revolutionise existing methods – that have the biggest impact.

“As we move forward, it’s these evolutionary advancements, rather than revolutionary disruptions, that will continue over time to bring the constant change.

“As logisticians, until we all have access to a Tardis, we have to work with the ever-evolving tool box we have. AI is now part of that tool kit, but will it become an everyday tool or become that specialist tool used with power, but not very often? Time will tell.”

For more information on Prological, click here

Send this to a friend