How Generative AI is Transforming Supply Chains - Top 3 Real World Examples from DHL, Walmart & C3

Summary

Supply-chain veteran Jamal Hai introduces a two-part video series sparked by audience interest in AI applications. He argues AI isn’t new to the industry—breaking it into perception (OCR, telematics), physical (warehouse robots), generative (dynamic routing, inventory optimization, procurement automation) and agentic AI—then showcases real-world examples: DHL’s dynamic last-mile routing, Walmart’s demand and inventory optimization, and C3.ai/Unilever-style procurement automation. He warns that 85% of AI projects fail (Gartner) due to unclear goals, poor leadership, bad data, weak change management, and fraud. Key takeaways: insist on human judgment, cross-functional collaboration, high-quality data, and thoughtful people change management.

Takeaways

• 😀 AI has been in use in supply chain for a long time, starting with perception AI like document recognition and temperature control in trucks.
• 😀 Physical AI in supply chain includes robotics, such as Amazon's use of Kiva robots for efficient product picking and delivery.
• 😀 Generative AI, like ChatGPT, is making waves in supply chain by improving dynamic routing, inventory optimization, procurement automation, and more.
• 😀 Agentic AI, such as AI chatbots for customer service and dynamic sourcing, has seen more widespread adoption in supply chains since 2020.
• 😀 85% of AI initiatives in supply chain fail due to reasons like unclear objectives, poor leadership, unrealistic expectations, and lack of proper change management.
• 😀 Poor data quality is a major reason for AI failures—garbage in, garbage out—making accurate data essential for success in AI projects.
• 😀 Be cautious of AI scams, as the term 'AI' is often misused; ensure that solutions are truly AI-driven rather than relying on manual work in the background.
• 😀 DHL has leveraged generative AI for dynamic routing and last-mile delivery, optimizing delivery times and reducing fuel usage by adjusting routes in real time.
• 😀 Walmart uses AI for demand and inventory optimization, redistributing warehouse stock based on real-time data, improving service, and reducing waste.
• 😀 Procurement processes are being enhanced with AI, such as the C3 AI system that uses AI for spend analytics and supplier negotiations, reducing workload and improving cost efficiency.
• 😀 Key lessons for successful AI implementation in supply chain include the importance of human judgment, collaboration, data quality, and thorough change management, especially people management.

Q & A

Why is AI not really a new concept in supply chain?

AI has been used in supply chain for decades, with early applications like document recognition (e.g., optical character recognition) and temperature control in trucks. These technologies have helped automate tasks like scanning documents or monitoring transportation conditions, laying the foundation for more advanced AI applications today.

What are the four types of AI in supply chain discussed in the video?

The four types of AI discussed are: 1) Perception AI, which involves technologies like document recognition and temperature monitoring, 2) Physical AI, used in robotics for automating processes, 3) Generative AI, used for tasks like dynamic routing and inventory optimization, and 4) Agentic AI, which powers tools like automated customer service and sourcing.

What are some common reasons AI initiatives fail in supply chain?

According to Gartner, 85% of AI initiatives in supply chain fail. Common reasons include unclear objectives, incompetent leadership, unrealistic expectations, poor change management, and poor data quality. Additionally, scams in the AI space, where companies falsely claim AI capabilities, can lead to failed implementations.

Can you explain the concept of generative AI in the context of supply chain?

Generative AI refers to AI that can generate new content or solutions based on inputs, like ChatGPT. In supply chain, generative AI can optimize tasks like dynamic routing for deliveries, inventory management, and procurement processes by analyzing data such as traffic patterns, demand trends, or supplier performance.

What are some examples of generative AI in supply chain operations?

Three examples of generative AI in supply chain are: 1) DHL's dynamic routing and last mile logistics, where real-time data is used to optimize delivery routes, 2) Walmart's demand and inventory optimization, which uses data to improve stock levels and reduce waste, and 3) Procurement automation by C3 AI, which streamlines supplier negotiations and optimizes purchasing decisions.

How does AI help optimize last mile delivery, according to the DHL example?

DHL uses generative AI to optimize last mile delivery by adjusting delivery routes based on real-time data like traffic and weather patterns. This dynamic routing reduces fuel consumption, minimizes delays, and improves overall efficiency in delivery operations, leading to better customer satisfaction.

How does Walmart use AI for demand and inventory optimization?

Walmart uses AI to continuously track demand patterns and inventory levels across its many stores. The AI algorithm helps redistribute stock based on real-time data, balancing inventory levels and ensuring faster delivery, reduced waste, and improved operational efficiency.

What role does generative AI play in procurement, as demonstrated by C3 AI?

Generative AI in procurement, as shown by C3 AI, automates the analysis of procurement spending and supplier performance. It helps streamline the negotiation process by analyzing historical data and optimizing supplier negotiations, thus reducing the workload of procurement teams and improving overall efficiency.

What are the four key lessons for AI implementation in supply chain?

The four key lessons are: 1) Human judgment is crucial; AI alone cannot solve complex supply chain problems. 2) AI projects often require collaboration across different departments. 3) Quality data is essential for successful AI outcomes; garbage in, garbage out. 4) Change management, especially people change management, is critical to ensure smooth transitions when implementing AI in supply chain operations.

How does human judgment play a role in AI implementations, and why is it important?

Human judgment is essential in AI implementations because AI can produce outputs based on patterns in data, but these outputs need to be interpreted and validated by humans. Blindly following AI's suggestions can lead to errors or problems, as AI does not fully understand the broader context or nuances that humans can assess.