Introduction
Oksana Lewen once spent an afternoon translating the word “appointment” into three versions of German that all seemed correct and all failed at the counter. One was too formal, one too vague, one simply not how people spoke in that office.
Oksana lives in Zurich and works as an AI language coach for adults rebuilding their lives after displacement. Before that she taught literature. The classroom changed; the attention to words did not.
Story of the Path into AI
After arriving in Switzerland, Oksana saw how many people stumbled not over grammar alone, but over forms, dialects and the hidden rules of official language. A language model first became a translator, then a practice partner. It could simulate a conversation at a counter, but it also invented phrases that sounded fluent and were socially off.
She was dealing with professional recognition, a new language and the feeling of being a beginner again. Instead of hiding that, she turned it into teaching material. She compared automatic translations with everyday Swiss expressions, collected misunderstandings and wrote role-play prompts for offices, rental situations and school meetings.
Her first useful set explained administrative vocabulary and made learners practise questions before real appointments. It had to include a strict reminder: preparation is not legal advice, and uncertain cases need a qualified human contact.
Current Work
Today Oksana works in an integration programme and tests AI-supported learning paths for adults. In one exercise on rental contracts, learners first formulate their own questions, then compare them with AI suggestions and finally write down where to seek real advice. The model is a rehearsal room, not the authority.
Participants report that they prepare official conversations more calmly and ask more follow-up questions. Oksana considers that a practical success. She remains careful: AI can bridge language gaps, but it must not replace the institutions responsible for fair information.
Personal Advice
“AI can be a bridge, but you must know which side responsibility is on,” Oksana says. She tells learners to use the tool for practice, vocabulary and courage, while keeping important decisions anchored in people who are accountable.
Key Facts
Age and place: 45, Zurich.
Background: displacement, teaching, professional recognition.
Entry into AI: practice set for administrative vocabulary and appointment role-plays.
Focus today: integration and language learning.
Typical tools: language models, role-play prompts, translation comparison.
Werkstattnotiz
Oksana has a page titled “almost right.” It holds phrases that no dictionary would reject and no local person would naturally say. She reads the page before updating exercises. Her current concern is how to teach confidence without teaching people to sound more certain than their situation allows.