Thesis worker

Master Thesis: Agentic AI for Intelligent Logistics

Background

Logistics operations involve a continuous flow of decisions related to material availability, transport, packaging, suppliers and operational deviations. Many of these decisions are still coordinated manually across people, systems and organizational boundaries.

Scania Logistics is exploring how agentic AI can enable more autonomous logistics operations. Unlike traditional automation, an AI agent can gather information from multiple sources, evaluate alternatives, interact with systems and people, and take action within defined guardrails.

The ambition is not unrestricted automation. It is bounded autonomy, where humans define objectives, decision rights and constraints, while agents handle suitable operational decisions and escalate uncertain or high-impact situations.

Thesis objective

The objective of this thesis is to investigate how agentic AI can support or autonomously manage an operational logistics decision flow.

Together with Scania, the student will select a relevant logistics area based on business priority, available data and technical maturity at the start of the thesis. Potential areas may include inbound logistics, transport, packaging, supplier collaboration, material availability or deviation management.

Possible research questions

• Which operational logistics decisions are suitable for AI-assisted or autonomous execution?
• How should responsibilities between humans and AI agents be designed?
• What evidence, confidence levels and guardrails are required before an agent may act?
• How should an agent respond when information is incomplete, conflicting or uncertain?
• How can the solution be designed for reuse across several logistics flows?
• How should operational value, safety and autonomy be measured?

Expected work

The thesis is expected to include:

• Analysis of a selected operational logistics decision flow
• Identification of suitable agent responsibilities and autonomy boundaries
• Design and implementation of an agentic AI prototype
• Testing using relevant operational, historical or simulated scenarios
• Evaluation of business value, decision quality, risk and scalability
• Recommendations for further development and industrial implementation

The exact use case and technical scope will be defined together with the student when the thesis begins.

Desired background

We are looking for a student or student pair within areas such as artificial intelligence, computer science, industrial engineering, logistics, operations research or a related field.

Experience with AI agents, machine learning, optimization, software development or simulation is beneficial. More importantly, we are looking for curiosity about how AI can redesign operational decision-making in a complex industrial environment.