Tech@Lunch April: Competence-aware Training & Instruction

Mismatched work instructions

Manufacturing companies often face the challenge of training and supporting a workforce with very diverse experience levels. Standard work instructions are generic: too detailed for experienced workers but not detailed enough for novices. This mismatch slows down learning, creates frustration, and increases the risk of mistakes.

Personalized work instructions through AI

In this talk, we show through examples how AI is used to personalize the instructions based on a worker's competencies and his proficiency level. Large Language Models (LLMs) automatically adjust the level of detail, while knowledge graphs provide domain expertise to make outputs consistent and reliable. Novices are guided step by step to prevent mistakes, while experienced workers receive short, to-the-point instructions that help them work faster. Built-in micro-learning—like quizzes and interactive chatbot Q&A—boosts knowledge retention and upskilling without interrupting production.
 
We will demonstrate how this works in practice, including integration with digital work instructions platforms like Azumuta, and discuss the impact for companies looking to improve productivity, quality, and workforce flexibility.