It is nothing new to use digital tools to assist supply chain management. For example, brand factories can establish situation analysis systems to simulate contingency plans for disasters, or add RFID (Radio Frequency Identification) tags to each batch of goods to control every aspect of logistics. information to ensure on-time delivery.
In the past, when companies invested in supply chain digital transformation, they often encountered data scattered everywhere, making it difficult for decision-makers to fully grasp the information and make immediate contingency plans. The second challenge is that the technical threshold is high. For example, if a Japanese factory encounters a problem and wants to assess the amount of losses and inquire about alternative production lines, the work is too complex to be handled manually, but the entry threshold for writing instructions in the program is also high.
Generative AI can aggregate unstructured data and build a supplier knowledge base for decision-making reference
Supply chain management applications of generative AI can be divided into four levels based on maturity.
Level 1: Solving specific tasks
Purchase commercially available AI tools to solve the pain points of a single task. For example, when contacting Korean customers, use ChatGPT to write Korean letters.
Level 2: Optimize workflow
Accelerate the efficiency of specific workflows. For example, the technology company Shipwell provides a logistics management system to help customers plan transportation routes and evaluate transportation costs. The biggest difference between Level 2 and Level 1 is that when enterprises introduce commercially available AI solutions, they will make some customizations based on internal processes.
Level 3: Changing Workflow
Think of generative AI as an agent or virtual robot that automatically breaks down complex tasks and executes them based on the user's goals. For example: when encountering a disconnection crisis and need to urgently think of alternatives, AI can analyze the root causes, simulate multiple solutions for enterprise reference, and proactively send letters to notify relevant units, which will change the existing workflow. Companies need to have the ability to develop AI solutions to reach this level.
Level 4: Cross-unit process automation
Multiple virtual robots operate simultaneously to automate cross-department processes. For example: the purchasing unit discovers that the cost of parts has increased through AI, and automatically notifies the R&D unit AI, which designs products with different parts to provide management with reference, thereby reducing procurement costs.
It is too difficult to achieve Level 4 at this stage. BCG assesses that the development focus of enterprises in the next 3 to 5 years is Level 3. Currently, most applications are focused on Level 1 and 2, that is, using AI to improve work efficiency when seeing a single pain point.
As companies progress to developing agent AI, menial tasks can be reduced to 30%
I believe that only by allowing generative AI to drive changes in workflow will corporate competitiveness be enhanced. Take the supply chain affected by natural disasters as an example. In the past, people would make calls to understand the situation, take inventory, find alternative suppliers, and convince customers to accept new solutions.
During this process, 60% of the time is spent on administrative communication tasks and only 40% is used for decision-making. By using AI to automatically generate supply chain contingency plans and inventory suitable suppliers, companies can reduce the time spent on trivial tasks such as strategic planning and inventory counting to 30% from 70%. A supply chain that responds quickly and makes careful decisions can create differentiated advantages.
Enterprises investing in AI can grasp the "10-20-70" principle, which means that if AI can create value, 10% of its contribution comes from algorithm models, 20% from digital technology and data collection, and 70% from people, processes and Organizational changes. In other words, companies must achieve the benefits of AI through effective management methods, education and training, and cultural changes.
