Walmart, Avon and the AI Agility Race
AI is increasing efficiency and resilience in the supply chain, but it is more challenging than it looks.
Right now, supply chain professionals across industries are grappling with the same question: "How will AI change the game for my organization, my role, and my future?"
The truth is that AI's impact is not a one-size-fits-all scenario. Some aspects of supply chain management may experience minor improvements, while others are on the brink of a complete overhaul.
A recent study surveyed over 1,700 supply chain management leaders worldwide and found that over half of companies use AI, automation, or machine learning for at least one supply chain management application.
However, implementing AI in supply chain management is not without its challenges. Lack of quality data, complex supply chain networks, and the need for specialized expertise are just a few of the hurdles companies face.
In today's newsletter, we'll explore:
Recent developments on how AI is used in the supply chain
Common challenges of implementing AI in the supply chain and solutions
How Aramex improves overall customer experience with AI
Let’s get started!
AI in supply chain: Recent developments and their impact
1. SPS Commerce applies AI to improve retail supply chains
SPS Commerce, a leading retail supply chain solutions provider, uses AI to analyze trading partner transaction data and generate demand forecasts, persona-based marketing, and other supply chain enhancements.
With 120,000 customers across 85 countries, including major retailers like W.W. Grainger Inc., The Home Depot Inc., and Target Corp., SPS Commerce is well-positioned to drive innovation in the retail supply chain.
By applying AI to its vast data repository, SPS Commerce has achieved efficiency, speed, and accuracy levels ten times higher than non-AI methods for tasks such as demand forecasting.
The company is also using AI to expedite the onboarding of supplier product data, create customer personas for targeted marketing, and more.
This could lead retailers to make more informed decisions, improve customer experiences, and optimize their supply chain operations.
2. Walmart uses AI for supply chain resilience
Retail giant Walmart uses AI, machine learning, and vast computing power to transform its demand forecasting, inventory flow, and cost optimization approach.
According to Parvez Musani, SVP of End-to-End Fulfillment at Walmart U.S. Omni Platforms and Tech, the integration of these technologies has changed how the company approaches supply chain processes, from understanding customer needs at a granular level to simulating potential disruptions and preparing proactive responses.
Walmart's AI-driven supply chain management has enabled the company to maintain agility, meet customer expectations, and navigate the challenges posed by unexpected events and disruptions.
By automating decision-making and investing in a reliable, reconfigurable supply chain, Walmart is setting an example to manage supply chain risks efficiently.
3. Avon partners with Blue Yonder to streamline demand planning
Beauty company Avon International has partnered with Blue Yonder to establish a centralized "Planning Hub" for demand and supply planning. Using Blue Yonder's AI and machine learning solutions, Avon aims to reduce inventory, improve forecasting accuracy, and respond quickly to market changes.
Blue Yonder's solutions offer the scalability of machine learning forecasting and the extensibility and explainability of causal factors such as campaigns, price, promotions, catalog position, and seasonality.
These factors train the ML models to predict future demand, enabling Avon to make more informed decisions and improve overall supply chain efficiency.
The success of this collaboration could inspire other companies to use AI-driven demand planning solutions for more efficient and sustainable supply chains.
4. RobobAI uses AI to optimize spend analysis and procurement
RobobAI, a global FinTech company, uses AI to help organizations streamline their supply chain management.
By applying AI to vast amounts of spending data, RobobAI's platform offers visibility and insights, allowing companies to reduce costs and improve operational efficiency.
According to Nitin Upadhyay, the company's Chief Data and Innovation Officer, a typical organization spending $1 billion annually on goods and services can realize savings of up to $6-8 million per year by adopting RobobAI's AI-driven insights.
These savings come from various areas, including operations optimization, payment restructuring, and contract expansion opportunities.
Common challenges of implementing AI in supply chain management and solutions
Implementing AI in supply chain management has its challenges. Organizations often face hurdles in scaling AI solutions effectively.
However, you can overcome these obstacles with the right strategies and tools. Let's explore some of these challenges and potential solutions.
1. Training costs
Current AI models for supply chain management often require extensive datasets and substantial computational resources, which can lead to high training costs.
Potential solutions:
Start with pre-trained models on general supply chain data and fine-tune them for your specific use case.
Use cloud-based AI platforms like Amazon SageMaker or Google Cloud AI Platform for cost-effective and scalable model training resources.
Optimize AI training costs based on project needs and budget constraints using flexible cloud solutions.
2. Operational costs
You'll face expenses for maintaining and updating AI systems, including software licenses, cloud computing resources, and potential hardware upgrades.
Additionally, you may need to hire or train specialized personnel to manage and interpret AI outputs, adding to your labor costs.
Potential solutions:
Use serverless computing platforms like AWS Lambda or Google Cloud Functions to reduce infrastructure management costs.
Adopt automated model monitoring tools such as Amazon SageMaker Model Monitor or DataRobot MLOps to quickly identify and address performance issues.
Implement containerization with Docker and orchestration with Kubernetes to improve resource utilization and scalability.
3. Complexity
Supply chains are inherently complex, with numerous interconnected processes, stakeholders, and variables. Implementing AI in this environment presents significant challenges.
You'll need to ensure that your AI models can handle the intricacies of your supply chain, including multiple suppliers, various transportation modes, and diverse customer demands.
Potential solutions:
Implement digital twin technology using platforms like IBM Watson IoT or Microsoft Azure Digital Twins to model and simulate your supply chain.
Use process mining tools like Celonis or UiPath Process Mining to identify inefficiencies and optimization opportunities.
Deploy a supply chain control tower solution like Blue Yonder or o9 Solutions to gain end-to-end visibility and decision-making capabilities.
Use graph database technologies like Neo4j or Amazon Neptune to model and analyze complex supply chain relationships.
4. Scalability
As your business grows and evolves, your AI solutions must scale accordingly. You'll face challenges ensuring that your AI models and infrastructure can handle increasing data volumes, more complex supply chain networks, and changing business requirements.
Potential solutions:
Use cloud-native technologies and microservices architecture using platforms like Amazon ECS or Google Kubernetes Engine.
Implement data streaming solutions such as Apache Kafka or Amazon Kinesis to handle large volumes of real-time data.
Adopt MLOps practices and tools like MLflow or Kubeflow to streamline model deployment and management at scale.
5. Data quality and availability
You may encounter challenges in collecting, cleaning, and integrating data from various sources across your supply chain. Issues such as incomplete, inaccurate, or inconsistent data can significantly impact the performance of your AI models.
Potential solutions:
Implement data quality management tools like Talend or Informatica to automate data cleansing and validation.
Use data integration platforms like MuleSoft or Dell Boomi to connect disparate data sources across your supply chain.
Deploy blockchain solutions like IBM Blockchain or Hyperledger Fabric to share secure and transparent data with supply chain partners.
Utilize data labeling platforms like Scale AI or Appen to improve the quality of training data for your AI models.
How Aramex improves overall customer experience with AI
Aramex, a global logistics and transportation solutions provider, faces a unique challenge in emerging markets: the lack of standardized postal codes and reliable address systems.
This makes last-mile deliveries inefficient and prone to errors, as drivers heavily rely on descriptive addresses.
To solve this problem, Aramex has developed an innovative solution using AWS AI services, particularly Amazon SageMaker.
The company deployed a generative machine learning algorithm that uses text-matching to convert descriptive addresses into precise XY coordinates, reducing the need for manual intervention.
The solution architecture consists of two main flows:
Advanced analytics and machine learning:
Amazon Glue ingests data from various sources into Amazon Redshift and Amazon S3.
Data scientists use this data to build, train, and optimize machine learning models.
Empowering end users:
Amazon SageMaker fetches data from Amazon Redshift and Amazon S3 to train and optimize models.
Once the models achieve the desired precision, they are deployed onto Aramex's operating systems and driver apps using Amazon API Gateway and AWS Lambda.
These models provide accurate address geocoding and XY coordinates in real time.
The benefits:
Early address geocoding allows Aramex to optimize last-mile capacity and plan routes more effectively.
Average daily success rates have improved by 10% across core and emerging markets.
Driver productivity has increased by 10% in core markets due to enhanced navigation tools.
Aramex has transformed its address management process in emerging markets, improving operational efficiency and customer experience.