The Future of Big Data & AI in Logistics
Historically supply chains have produced large quantities of high value data; optimising this data, analysing it, and learning from it, is a challenge many organisations face. By harnessing Big Data and Artificial Intelligence (AI) to manage supply chain functions, specifically logistics in this case, organisations can forecast demand, increase accuracy, better understand buying cycles, estimate future warehouse capacity, and solve complex operational challenges.
In this roundtable we look to the future of Big Data and AI in logistics with:
JB: Jennifer Bisceglie, CEO & founder of Interos
MB: Mike Bhaskaran, COO, Logistics & Technology at DP World
DS: Drew Sonden, EMEA Product Lead, Blue Prism
In the past five years - How has the logistics industry changed? What does the current landscape look like as 2021 comes to a close?
JB: The current industry landscape for logistics has become one of interconnectivity and complexity. Globalisation and digitalisation have acted as connecting forces upon enterprise supply chains, intertwining industries and the organisations that work within them. While worldwide connectivity has its benefits, it does also mean that supply chain resilience is made more fragile as disruption events can ripple through the web of thousands of organisations.
One example of this is the semiconductor – or ‘computer chip – supply chain, which is right now highly disrupted. The semiconductor supply chain is extremely complex and globally interconnected, with the production of a single computer chip often requiring more than 1,000 steps passing through international borders over 70 times. This interconnectivity and globalisation of the supply chain has exacerbated disruption borne from the pandemic and has led to shortages of semiconductors in UK industry, hampering important sectors such as auto production.
MB: Global logistics has become complex due to Covid-19, causing urgency and bottlenecks throughout the supply chain. For example, the Suez Canal situation was made worse due to existing industry constraints. There have been further supply chain disruptions, such as lack of container availability and rising ocean freight costs, that have had a sudden and unexpected impact on logistics.
To promptly manage urgent merchandise exceptions when there are disruptions, visibility and transparency become imperative and real-time solutions are needed to provide quality insights. Today, there are still significant hurdles that can’t be ignored. To overcome these, DP World has derived a long-term digital technology plan through SeaRates, a freight-rate spot marketplace, married with CARGOES Flow, the enterprise tracking tool for intermodal shipments. What these solutions provide is supply chain visibility with a real-time optimisation engine that enables alternative routes, mitigates delays as well as monetary risks.
Longer supply chains face additional challenges beyond delays, such as working capital that limits cargo owners ability to move quickly and at scale. This is inherently true of cargo owners in developing economies that struggle to find the funds to operate a business with fluidity. Hence, another product we launched was CARGOES Finance. The objective, to enhance business for importers, exporters and logistics companies around the globe by providing access to financing for receivables and payables.
DS: It’s no secret that good logistics is underpinned by data. The ability to effectively forecast demand, manage supply, manufacture and distribute effectively, all this relies on strong and reliable historical and live data, with the most successful companies the ones who are able to leverage data on one part of the supply chain to impact others. It can be something as simple as ordering logistics capacity based on production volumes, or more complex analysis like adjusting product lines to account for variables like weather or political forecasts.
It’s perhaps ironic then, that over the past few years, the biggest change in big data is that companies are becoming more circumspect about its potential. This is understandable when we consider the claims that some vendors were making with respect to the transformational potential of big data platforms.
From an AI perspective we’ve seen organisations shift to a much more discrete application of AI solutions, applying the technology within narrow bounds that offer proven results. For example, we have one customer, a US logistics firm which uses AI decisioning in conjunction with Blue Prism to deliver more effective vehicle maintenance. Our intelligent automation platform gathers IoT sensor data from their fleet of over 14,000 vehicles and feeds this to an AI that identifies maintenance issues. This allows Blue Prism to generate personalised maintenance schedules for each vehicle – and to send live alerts to drivers of trucks in danger of critical failure.
We have several firms that combine NLU with Blue Prism to more effectively manage customer engagement. Customers engage via text channels such as email, the digital workforce picks these messages up and shares them with the AI toolsets. These identify sentiment and intent from the messages, allowing Blue Prism to triage them, pass them to the appropriate team or, where possible, to automatically resolve the customers’ issues.
Both of these examples reference the third big change that we’ve seen in the market which is that businesses are becoming ever better at using technologies like AI and big data to augment the power of the core Blue Prism intelligent automation platform. Whilst technology is a long way from being able to make subjective decisions, AI and big data allow us to apply a lot more nuance to logical business processes and meaning that intelligent automation can do more to free up the time, effort, and capacity of the human workforce.
How can big data and AI help supply chain organisations become more resilient?
JB: AI-powered technologies and big data can help organisations build true operational resilience in their supply chains. For example, with the right signals and data sources, organisations can anticipate labour shortage trends and preemptively evaluate the benefits of increasing inventory where needed. Furthermore, AI can help enable more seamless inventory management. For example, a manufacturer with geographically diverse suppliers, who monitors and identifies growing case counts in one region, could shift order commitments to a less impacted region while building safety stock. With the right data, mapping and monitoring, the supply chain team can evaluate conditions faster and move more quickly to lessen costly impacts. The current disrupters to the UK’s supply chain are complex, but this complexity can be managed for organisations with smart operational resilience solutions that make use of AI and big data.
MB: Data is crucial. Containers have a vast volume of products, and commonly, delays can occur throughout a customer’s supply chain. Data provides solutions and insights. Take, for example, a customer waiting for its shipment to arrive, assuming it will be on time. The reality is that supply chains are unpredictable with variables that can be unforeseen, as we all now know. Big data, AI - and I’ll also add document digitisation leveraging blockchain - can provide customers with supply chain visibility and velocity, enabling a procurement manager, for example, with the opportunity to go and find new alternatives. From a visibility perspective, it could be the decisive action required to optimise supplier or route. From a velocity perspective, digitising signatures, identification, and documentation processes from liner to gate can accelerate a container through a port, where data becomes interoperable through various stakeholder systems.
All things being equal, the idea is to reduce delays and port congestion by digitisation that provide insights and actions to be taken by the responsible parties from port communities to customs and beyond - in real-time and predictively.
We heavily invest into our stackability matrix strategy through ports, which is an AI exercise consisting of looking at how containers are stacked, what potential damage could occur, safety and efficiency of movements. For example, making sure heavy merchandise is not added on top of lighter ones, to reduce the number of moves a container makes through the port. We do this through AI, and it becomes crucial for just in time customers that look to limit on-hand supply and materials to store, but request as needed.
DS: Even restricted to the limited definitions that I outlined, big data and AI are both incredibly powerful capabilities, however they are predominantly limited to generating insight. A big data platform may be able to highlight that, counter-intuitively, a business should reduce production of a particular product because it performs better in the market when it is viewed as exclusive by consumers. An AI computer vision engine may be able to highlight defects in components invisible to the human eye. However, these insights are useless unless they are acted upon. With intelligent automation, we can not only commission insights without human instruction, but we can also act appropriately on the results: reducing supplier orders and slowing down the production line in the first case; automatically routing defective components to a recycling unit so they can be dismantled and the materials effectively reused.
The power not just to generate novel insights, but to act on them in real time gives us huge scope to effectively navigate an increasingly fragmented and complex supply chain landscape. It enables leaders to achieve greater operational productivity, agility, and resilience. This is more relevant than ever today, as we see businesses facing huge challenges, not just with the worrying and ongoing lack of logistics capacity, but also shortages in a range of crucial raw materials such as silicon chips and sheet aluminium.
Ultimately, AI, big data, and Intelligent Automation can all deliver multiple layers of value when applied to a single challenge or business issue, but it is when they are applied within the context of a top-down digital transformation programme that we really see the most significant benefits. As the market matures, I’m starting to work with an ever larger number of customers who are automating by exception – mapping out their end-to-end supply chain processes and asking not ‘what can I automate’, but instead starting out with an automation-first mentality and ruling out those activities that are truly subjective, or rely on human-to-human engagement.
As we look towards 2022 - How do you see big data and AI evolving?
JB: As the adoption and popularity of AI/big data solutions continues to increase and more businesses start to see the benefits – smarter logistics, faster responses to emerging supplier risk, and a deeper knowledge of their extended supply chain – the next front will be predictive analytics and true digital twins, accurate digital models of organisations’ entire extended supply chains that can enable companies to game out various risk scenarios. These could include cyber-attacks or labour shortages impacting critical sub-tier suppliers deeply embedded within the supply chain and would enable businesses to pre-solve problems before they happen. As AI and big data solutions begin to dissolve the barriers that keep supply chain information siloed, such as widely differing formats for supply chain data, this level of predictive insight will become much more feasible.
MB: Procurement, business-to-business commerce and container stacking will be key in 2022. For example, I see African coffee farmers identifying new roasters in different markets from leveraging big data to find these opportunities. Through Dubuy, a B2B e-commerce platform, we’re providing new market access and matchmaking opportunities for suppliers of goods, commodities, food items, clothing, and technology.
Solving stacking challenges will be a critical part of solving many supply chain challenges with automated infrastructure and AI at the helm. Our Boxbay system is a good example of this. A standard port, for example, can stack four to five containers high depending on the load. Boxbay can stack 11 containers high. This significantly reduces the space required to manage a port and the time for a haulier to come and drop off or pick up the container as they will not have to wait for other containers to be moved around.
DS: Up until recently, the ability for organisations to leverage AI and big data has relied on the recruitment of specialist technical resources such as data scientists. The transformational capabilities of the tool sets were without question, but deploying them effectively required significant time, resource, and investment, meaning that only the largest or most focussed businesses were able to take advantage.
What we are beginning to see is a marketplace that offers highly capable AI and big data tool sets that are accessible with a much lower technical threshold. From a supply chain perspective, this offers us the capability to undertake much more complex forecasting activities, which seamlessly link variables from each step of the process from availability of raw materials, logistics capacity, and demand impactors. big data allows us to collate and present the data required, whilst AI tools offer the capability to make nuanced predictions based on a variety of scenarios.
There is no doubt that the capabilities of AI and the benefits they can bring to organisations will continue to advance. As digital workers become smarter, we see human workforces and their digital counterparts become more and more intertwined. The future workforce will be a seamless and equal blend of digital workers, human employees, and existing systems, all working collaboratively in a way never seen before. These new unified workforces will create greater employee and shareholder satisfaction, whilst, most importantly, providing customers with even better experiences.