Most language schools sell private lessons because that is what students think they want. One teacher, one student, full attention. It feels like the premium option, so it gets priced and marketed like one. After ten years of teaching, I am convinced it is the wrong default for the thing people are actually buying.
Because what people are buying is not grammar. It is not vocabulary. You can get both from an app for free. What people are buying is the ability to open their mouth in a real conversation and not freeze. That is confidence, and confidence is a social skill. You do not build a social skill in a room with one person whose entire job is to be patient with you.
What a decade of lessons actually showed me
Speaking with a tutor is not the same as speaking with peers, and the differences are the whole point. With a tutor, the social pressure is off — they will wait, they will not judge, they have heard every mistake a thousand times. That safety is useful for a while. But it is also artificial. Nobody in your real life is a trained, infinitely patient listener. Your colleague in a meeting is not. The officer at the visa interview is not. The nurse manager on your first shift is not.
The register is different too. One-on-one with a teacher, students drift into a careful, corrected, classroom English. In a group, they have to do the real work: hold the floor, jump back in after someone interrupts, disagree without sounding rude, recover when they lose the thread. Those are the exact moments students tell me they dread — and they are precisely the moments a private lesson never rehearses.
And there is one thing a group gives you that no tutor can: you get to watch someone at your level try, stumble, and recover. That is where the belief comes from — if they can do it, so can I. A teacher succeeding in front of you proves nothing. A peer succeeding in front of you changes what you think is possible.
So why does everyone still sell 1-on-1s?
Because groups are genuinely hard to run well, and most schools never solved the two problems that make them fail.
The first is airtime. Put six learners on a call and, left alone, two confident ones take eighty percent of the talking while the quiet four hide with relief. Everyone leaves feeling like they attended. Only two of them practised.
The second is momentum. Group energy dies when the teacher becomes a bottleneck — correcting one person while five wait, or steering the whole hour so tightly that no real conversation ever breaks out.
Our answer to both is an AI facilitator that sits inside the session doing one narrow job well: managing turns and holding the thread. It is not a teacher. It is a turn-taking referee with a memory. It notices who has not spoken and hands them the floor. It remembers what you said last week and calls it back this week. It keeps the tutor free to coach the moment instead of policing the clock. The teaching is still human. The traffic control is not.
Three versions we got wrong first
None of this arrived clean. The first cohort was too big — eight people — and the quiet ones simply disappeared behind the loud ones. We had recreated the exact failure we were trying to fix, just with more witnesses.
The second version fixed the size but had no facilitation layer, and one strong speaker quietly dominated every session. Nobody was rude. It just happened, the way it always happens, and the others slowly checked out.
The third version over-corrected: a rigid, minute-by-minute curriculum that guaranteed everyone spoke — and guaranteed none of it felt like a real conversation. Students hit their speaking targets and still sounded like they were reading a script, because they were.
What we kept is the narrow middle: small groups of four to six, a light structure that bends around the room, a human coach for the teaching, and the facilitator holding turns so no voice gets lost and no lesson gets scripted to death.
What the founding cohort is testing now
We are not done. Right now the founding cohort is helping us pressure-test the parts we are least sure of: the ideal group size for a mixed-level room, how much structure a session needs before it stops feeling like a conversation, and whether the confidence students build with peers transfers to the high-stakes moment they actually signed up for — the interview, the ward round, the pitch. That last one is the only measure that matters, and it is the one we watch most closely.
The honest version of the bet is this: private lessons are easier to sell and easier to run, and they under-deliver on the one outcome people are paying for. Small groups are harder to run and, done right, they are where the confidence actually gets built. We would rather do the hard thing that works.
— Deborah, founder