gensai.aiConversational interface
Oral English level assessment by AI
Project
For an English training center, an MVP was developed to offer a personalized voice-based English test. The learner talks for about 5 minutes with an AI-based conversational interface. At the end of the discussion, an analysis of the English level is generated, along with recommendations to guide the learner toward the most suitable courses.
The need
Assessing a spoken English level requires a real conversation, which a written questionnaire cannot provide: neither fluency nor comprehension in context can be measured that way.
How it works
The learner talks with the interface for about five minutes, as in a conversation. The voice exchange takes place in real time, with LiveKit and a real-time voice API. At the end, a level analysis is generated with course recommendations.
What the project shows
First designed as an MVP, the test made it possible to check the value of a conversation with an AI before going further. The same approach works for any structured oral interview, from assessment to practice.