Est. reading time: 7 minutes
Author: Steph Locke
Logistics businesses and logistics have a huge opportunity to unlock additional value in their business through the use of data and AI integration. Areas of improvement include streamlining operations, managing the supply chain, pricing dynamically, agile planning, and enhanced forecasting.
Logistics is one of the many industries and departments being heavily influenced by AI. Managing the flow of goods between locations, logistics companies operate within a complicated network of suppliers and customers, resulting in a range of tasks that can be easily automated and benefit from AI techniques.
Logistics firms have a lot to gain, with a reported
$1.3 trillion to $2 trillion per year in economic value due to AI integration into manufacturing processes and supply chains 1.
Success with AI is driven by data and data is now one of the strongest assets that any logistics company has available to them. However, while many have a lot of data at their disposal that they are not using for several reasons: lack of expertise, lack of data availability, data quality issues. Not having the data available in one central location and using the wrong tools to analyse are also key barriers.
In this article, I explore how logistics businesses that have delivered a huge opportunity to unlock additional value in their business through data and AI integration, including work with our own customer Freight Logistics Solutions.
Streamlining operations means using real-time data and alerts to optimise delivery routes, monitor performance, and respond to delays or issues as they happen. Route optimisation allows for on time deliveries at the lowest possible cost, but it requires real-time analysis of multiple data points.
AI techniques are perfect for this as they can analyse large amounts of data and continually learn from them. This allows the AI model to serve up the most efficient route in terms of cost and time – delivering a far more optimised process.
Shipping giant Maersk streamline IT operations and optimized the value of its IT resources by adopted Microsoft Azure. They migrated key workloads to the cloud, and modernized its open-source software, which included the adoption of Kubernetes on Azure. 3
Eddie Stobart support customers' preferences and constraints for integrating and communicating among our different systems. The result was cutting warehouse integration time from 26 weeks in half. 4
Freight Logistics Solutions (FLS)
FLS used to spend half their time adding data to spreadsheets to get around software limitations. We built a smart data warehouse to extract information and provide a real-time view to both staff and customers saving nearly 50% of staff time from this tedious task. Read the case study
Supply chain management
To remain competitive logistics companies must respond to customer demand and deliver value to users, using data to optimise supply chain management. There are lots of ways to optimise the supply chain such as real-time tracking.
Damco uses a Supply Chain Management system built on Microsoft Azure to provide precise tracking services to their enterprise customers ordering and receiving packages. 5
COVID 19, ambiguity, and price fluctuations have made dynamic pricing more valuable and have increased the importance of dynamic pricing. McKinsey 2 reports that logistics companies that transform their pricing could
increase revenue by 2 to 4 percent, translating to as much as a 30 to 60 percent increase in operating profit.
However, while pricing smart is important most companies don’t realise a demand surge is taking place until 30-50% of availability has been snapped up. Dynamic Pricing is an automated pricing strategy that is driven by demand changes in the market. It uses machine learning algorithms to analyse historical and event data to predict future demand. For logistics companies can adjust pricing in real time and gain from using demand intelligence.
DHL uses an integrated system with reliable and consistent data (Oracle central database) able to manage customer accounts, shipments, tariffs, and costs for all countries in the network. 6
AI tools can help logistics businesses analyse real-time data so that they can update their demand forecasting and supply planning. The result is an optimised chain flow and a reduction in the amount of waste. AI powered methods also reduce error rates significantly compared to traditional methods such as ARIMA, AutoRegressive Integrated Moving Average, and exponential smoothing methods.
Dynamic planning not only prevents stockouts, but local warehouses can reduce the holding costs.
ThyssenKrupp AG powered by Microsoft Azure, helps the company analyse and process more than two million orders per year and better serve its 250,000 global customers. 7
Predictive analytics remains the most important AI application within logistics given the abundance of supply chain data from which to draw predictive insights. There have also been significant increases in ecommerce making last mile delivery even more complex and the need for extracting better data intelligence even more vital.
Kuehne+Nagel’s use a unified data model to merge your data with external data, to show how to apply predictive analytics to identify and mitigate potential risks. 8
Autonomous devices work without human interaction with the help of AI. This includes self-driving vehicles, drones, and robotics. There has been a significant increase in autonomous devices in the logistics industry due to the industry’s suitability for AI. It’s deemed to be one of the most disruptive applications in the logistics industry.
DB Schenker uses visual technologies and logistics robots to alleviate some of the pressures stemming from the labour shortage in logistics, which has been widely reported by European businesses. 9
Data is one of a business' most important assets, but it can have a huge maintenance cost associated with it. DataOps is a set of practices and processes focussed on improving how we manage data to reduce the costs of supporting and using it.
Common goals for DataOps are:
- Reducing data integration latency to make decisions faster
- Reduce data quality defects to reduce the frequency of faulty decisions
- Reduce process implementation and maintenance times to ensure more informed decisions
- Reduce the complexity of individual operations to lower processing and storage costs.
The current best practice implementation of a data tier to support these goals is a “modern data warehouse”, combining a file-based data lake as an integration point and curated relational datasets to surface data for analysis and reporting. Getting this data infrastructure right is critical for helping get the most out of real-time data to optimise processes.
Machine Learning Operations (MLOps) focus on improving the speed and quality of the delivery of machine learning (ML) models into a production setting. Common goals for building effective MLOps into data science and machine learning engineering are:
- Reduce time to build and productionise models to lower the unit costs per model
- Improve model reproducibility and interpretability to decrease compliance costs
- Manage and monitor models at scale to ensure more time is spent on innovation
- Improve model quality to achieve higher ROI.
MLOps typically involves a mix of on-demand compute environments, orchestrated machine learning pipelines, version control, and Docker containers. This will be important if you start to build your own AI solutions, as it ensures quality solutions are delivered quickly.
These are lots of opportunities to use data and cognitive AI processes to improve your business. It’s important to think about adopting technology in light of your priorities and where you can unlock vital staff time. We recommend you start upskilling your team to be more aware of the potential technological solutions so you can foster innovation internally.
We also conduct Strategic Digital Reviews for manufacturing customers, who want to take stock of where they are and get expert advice on what technologies they should invest in and what areas they should upskill in. These reviews deliver actionable recommendations and are a great way to evaluate current operational processes and systems and identify gaps and improvement opportunities in a short space of time. Our Review can be 100% funded by the Enterprise Ireland Digitalisation Voucher initiative.
- How AI is spreading through the supply chain
- Getting the price right in logistics
- Maersk uses cloud to spur development of containerized solutions built on Kubernetes
- Integration Delivers On time for Eddie Stobart
- DevOps: Bringing logistics into the future and your orders to your doorstep
- DHL Express Successfully Implements Open Pricer Software to Support its Global Pricing Worldwide
- thyssenkrupp Materials Services ‘keeps calm and carries on’ – with its new ‘alfred’ AI solution to optimize its logistics network
- Integrated Logistics - Supply Chain Management 4PL
- DB Schenker case study: Autonomous robots in supply chain