What if the English you asked AI at work became your own personalized quiz? LingoQ: A Study on a Customized English Quiz App

LingoQ automatically converts English expressions that users ask about and correct with LLM-based AI during work into learning content. It consists of the desktop assistant LingoQuery, a quiz generation pipeline, and the mobile review app LingoQuiz, and creates TOEIC-style questions that reflect actual work contexts.

Users perform tasks such as looking up word/phrase meanings, translation, and proofreading in LingoQuery, and can send selected text along with screenshots to preserve the work context. The system then selects English-related Q&As and creates multiple-choice fill-in-the-blank questions, verifying answer validity and difficulty with AI and automatically revising up to 2 times.

A 3-week user study involving 28 participants confirmed sustained usage and learning effectiveness. Participants used LingoQuery on most weekdays, leaving an average of about 120 English questions per person, and LingoQuiz saw an average of 1 session per day.

The learning effects were also clear. Overall QESE scores rose by an average of 9.5%, and the reading and writing subscales also improved significantly. In terms of English proficiency, participants at the CEFR A level in particular showed an average increase of 4 points (out of 30), and participants rated the approach as more work-relevant and practically helpful than existing methods.

In the pre-study survey, 33 out of 49 respondents said they needed to study English for work, but less than half actually studied separately. On the other hand, 94% were using LLM-based AI for English at work, and the research team believes that turning this dependence into a learning opportunity can naturally connect work and learning.