👤 Persona

The Paper Was Accepted. One Question Undid the English.

A researcher’s English grows into an odd shape. You have read thousands of papers, and your writing is precise enough to survive peer review. Then the talk ends, one hand goes up, and none of that accumulation helps. You compose the answer in your first language, move it into English, check the grammar, and three seconds are gone; the chair moves to the next question. The problem is not your level, it is the route. A circuit that runs through translation cannot carry a live conversation. This is where Say Global English approaches researchers differently. First, the material is different. Instead of example sentences from a conversation textbook, you speak about your own research: your experiment, your data, your limitations. The concepts are already complete in your head, so English only has to be fitted over them. Second, the situations are different. In 1:1 sessions with six AI tutors you rehearse the talk, take anticipated questions, hold the corridor conversation after a session, run the video call with collaborators, and think through a reply to a reviewer. The Q&A work in particular is not memorising answers; it is building the moves people actually use at conferences, such as reflecting a question back to buy a second, or saying you do not know and continuing anyway. Third, the timing is different. It is open through the night an experiment is running and at dawn before a conference, and 25 minutes at a time is enough. Pronunciation and phrasing come back as immediate feedback, and progress, scores and streaks live on your account, so what you started on the lab computer continues on your phone at the venue. The interface and explanations exist in eight languages, so you never have to fight English merely to understand the instructions. It follows the principle 26 years of classrooms confirmed: a memorised sentence collapses in front of a question, while an acquired one keeps going even when it wobbles.