MSc · Aalborg University Business School
Two years in English. You learn to build things with data and AI, and to judge whether a company should use them.
Roman Jurowetzki · Associate Professor · 9 November 2026
Questions, any time.
Type or speak, Danish or English.
Anonymous.
01 Who's in the room
02 Where graduates work
work in North Jutland. Most graduates stay, and they don't have to: the rest are in Copenhagen, across Denmark and abroad.
02 Who hires them
02 What they do
The others take it into finance, strategy, consulting and companies of their own.
Real job titles
03 What you learn
Titles as in the 2026/27 curriculum. Programming, machine learning and language models, always applied to a business problem.
03 What it adds up to
Work with data and code. Produce a prototype that runs.
Does it work? How well, and at what cost?
Fit it to how the company works, to the law, and to ethics.
04 How it's taught
The Aalborg model, since 1974: a group, a real problem you formulate yourselves, a supervisor, an oral exam.
04 Real problems, real data
for a ship supply company, sanctions checks included
tracking the process in real time
for a restaurant in Aalborg, from bookings, weather and holidays
tested on bank annual reports
from financial ratios and the text of annual reports
from skills and AI exposure, deployed as an app
first place in an international shared task
studied from inside the student consultancy they founded
05 The AI angle
BDS is designed. GPT-3 is a research preview that few people have tried.
ChatGPT. Text, documents and images become data a business can use.
Building on foundation models instead of training your own. Teaching moves to evaluation and operation.
Coding agents write routine code. You still explain every line at an oral exam.
05 The AI angle
The tools you learn here will be outdated before you graduate.
What lasts is being the person in the room who understands what these systems can and can't do, and when to trust them.
06 The question you're asking
“Am I studying something that's about to be automated?”
06 Tasks and goals
Studies of AI exposure count tasks. About 80% of US workers have at least a tenth of their tasks exposed to language models.
Cheaper tasks don't remove the goal. Someone still decides what to do, and answers for the result.
Eloundou, Manning, Mishkin & Rock (2024), GPTs are GPTs, Science 384.
06 How organisations actually work
Before AI can do the work, someone has to turn that knowledge into data, rules and checks, and decide what counts as right. That is the job this programme trains you for.
And the parts that were never routine stay human: framing the problem, convincing people, the idea nobody asked for.
07 Your questions
Still open.
Later: roman@business.aau.dk
Study counsellors: ha-studievejl@business.aau.dk