The ibw – Institut für Bildungsforschung der Wirtschaft (Institute for Educational Research in Business) has been Austria's central institution for advancing and assuring quality in vocational education for five decades. It specializes in dual vocational training and supporting apprenticeships, trainers, and apprentices.
Through the lap.at platform, ibw provides high-quality learning materials — from study scripts to sample exams — for over 50 apprenticeship trades.
The Challenge: Oral Final Exams Can't Be Learned from Textbooks
Apprentices can prepare for written final exams using scripts and practice exercises. The oral exam conversation is different. It demands the ability to structure answers under pressure, apply technical terminology appropriately, and connect learning to their own workplace — skills that develop only through realistic practice situations. This sparked the idea to introduce an AI-powered exam simulator.
Two Formats, One Goal: Authentic Exam Simulation
During the pilot phase, ibw partnered with Trainexus to develop six scenarios for commercial and administrative apprenticeships according to Austrian exam standards. These consist of two formats:
With Document: Learners receive a real invoice or payroll statement in advance and are questioned directly about it by the AI examiner. The conversation focuses on document review (content, format, calculations), statutory requirements, and consequences of missing information.
Without Document: The conversation begins with the learner's workplace context — industry, company size, specific tasks — and develops along exam-relevant commercial topics: customer acquisition, revenue growth, business economics. This format particularly strengthens the ability to transfer theoretical knowledge to professional practice.
Scenarios were used by apprentices from diverse industries and company sizes. The AI examiner continuously adapts questions based on learner responses and probes deeper when explanations remain incomplete.
What the Simulator Achieves in Conversation
A typical dialogue in the document scenario shows how the simulator goes far beyond simple question-and-answer. When an apprentice mistakenly provides incorrect information instead of a company's VAT number for required invoice details over €400, the AI examiner corrects it factually and immediately asks a follow-up question: "Above what amount is the buyer's VAT ID mandatory — and why?" This forces active engagement, not passive knowledge repetition.
Personalized Feedback After Every Conversation
At the end of each session, Trainexus generates detailed, conversation-based feedback organized into five areas:
- Summary — What was the key development opportunity in this conversation?
- Strengths — Which answers and arguments already worked well, and why?
- Gaps with Guidance — Where is depth or precision missing? Each gap is addressed with a concrete three-step guide and sample phrasing for the exam.
- Mini-Drills — Short practice exercises (5–10 minutes) to target specific weak points.
- Action Plan — What should be done specifically before the next session?
Feedback doesn't just identify what was wrong — it explains why missing details cost points in the final exam.
Apprentice Feedback
"Apprentice feedback shows the simulator wins through its practical, exam-relevant exercises. They especially valued the opportunity to independently practice oral exam conversations with relevant topics. The personalized feedback based on actual conversation flow was also considered extremely helpful."
ibw – Institut für Bildungsforschung der Wirtschaft