LOG904-196 Agentic AI in Logistics and Supply Chain Management (Autumn 2026)

Agentic AI systems are rapidly changing how analytical, administrative, and decision-support work can be organized in logistics and supply chain management. This seminar gives students practical experience with using frontier agentic AI tools, such as Codex and Claude Code/Cowork, while also building their conceptual understanding of large language models, reasoning models, AI agents, and human-in-the-loop workflows.

The seminar is organized around a project-based learning format. Students will use agentic AI to design and build a working prototype, product, or analytical artifact that combines AI capabilities with traditional software or structured workflows. Example projects could include a customer service bot, a supply chain risk analyzer, an AI-supported procurement assistant, a logistics planning tool, or a decision-support dashboard.

Knowledge

- Explain the basic functioning of large language models, including tokens, prompting, context windows, model limitations, and the difference between language generation and reliable reasoning.

- Describe the role of reasoning models, tool use, memory, planning, and feedback loops in agentic AI systems.

- Understand how agentic AI can be combined with traditional software, data, and workflow design to support logistics and supply chain management tasks.

- Identify key risks and constraints in agentic AI use, including hallucination, data quality, privacy, evaluation, accountability, and over-automation.

Skills

- Use frontier agentic AI tools to plan, generate, test, debug, and document a small software-supported solution.

- Translate a logistics or supply chain problem into a feasible AI-supported project brief, including user needs, data requirements, workflow, and expected outputs.

- Build and iteratively improve a prototype, product, or analytical artifact using agentic AI in combination with conventional software components.

- Evaluate an AI-supported solution critically by testing outputs, identifying failure modes, and explaining the trade-offs of the chosen design.

General competence

- Work responsibly with agentic AI as a collaborator while maintaining human judgment, traceability, and critical oversight.

- Communicate the purpose, design, value, and limitations of an AI-supported logistics or supply chain solution to an informed audience.

- Reflect on how agentic AI may change professional work, coordination, and decision-making in logistics and supply chain management.

- Collaborate in a project-based setting where technical, organizational, and ethical considerations must be integrated.

Teaching consists of short lectures, live demonstrations, supervised hands-on exercises, project work, and peer discussion. The first part of the week introduces the foundations of large language models, reasoning models, agentic workflows, and practical use of frontier agentic AI tools. The remainder of the seminar is organized around student projects, where students develop an AI-supported logistics or supply chain solution with guidance from the lecturer.

The seminar will use a studio format: students work iteratively on their project artifacts, receive feedback during development, test their solutions, and prepare a final presentation and demonstration. The final day is reserved for oral project presentations and examination.

• Form of assessment: Submitted project artifact with oral presentation and demonstration. The submitted artifact and the oral presentation/demonstration together form the assessment basis
• Proportion: 100 %
• Duration: Project artifact submitted in Inspera before the final presentation. Oral presentation and demonstration on Friday during the seminar week. Exact submission deadline and presentation schedule to be confirmed with Molde.
• Grouping: Individually or in groups of 2-3 students. Individual grading is based on the submitted artifact, the presentation/demonstration, and the student's ability to explain and critically evaluate the solution
• Grading scale: Letter (A - F)
• Support material: All support material allowed for project development and presentation, provided that the submitted work clearly documents the use of AI tools, sources, and external materials.