What it really takes to create one lesson on StudyGo: behind the scenes

The moment a lesson is scheduled, whether by the auto-planner or manually through the tutor chat, preparation begins, and preparation is serious work.

“People often underestimate how long it takes to properly prepare a lesson,” says Lisa. “We really do have to read up on a subject, look up practice questions, and so on. It’s not like we can just answer an hour’s worth of questions on any topic without preparation.”

In practice, that means summarising an entire chapter, building a PowerPoint, finding the right practice questions, and anticipating what this student, at this level, actually needs to know for their specific test. Anna puts the time investment into perspective: what a tutor can distil into an hour of preparation would take a student roughly six hours to do on their own.

AI is part of the workflow. Tutors use it to handle large texts, get quick overviews of student questions, or generate challenging practice questions when StudyGo’s own content has been exhausted.

But the team is clear-eyed about its limits.

“AI gives you the answer, but not always in the right way,” says Anna. “It might provide a correct answer, but not in the format a teacher expects, or with the right keywords, or the right points for that specific exam.”

Stan, who leans on his background in applied psychology, puts it even more directly: “AI can hallucinate. I’ve seen a student submit a thesis with non-existent APA sources, all generated by ChatGPT. That’s when humans have to step in.”

Every lesson starts with a question: What do you actually need today?

A summary? Specific practice questions? A mock test before tomorrow’s exam? The answer shapes everything: the structure of the lesson, the pace, the examples used, even the language.

“You adapt to the level they’re actually at,” says Lannee. “Not the level they’re supposed to be at.”

With eight students in a group session, that’s a balancing act; one-on-one, it’s a conversation. Either way, it’s a skill AI doesn’t have, which is the ability to read a student, sense where they’re stuck, and meet them there.

Ask any tutor what they offer that AI doesn’t, and the answers come quickly.

“When I tell a student ‘I get it! French grammar makes no sense’ that actually helps them relax,” says Stan. “An AI can’t really do that.”

Anna adds: “Students sometimes assume we’re AI. They push for answers within a minute. But slowing down, taking time to think, giving a real explanation – that’s part of the value.”

There’s also something deeper at play. Tutors have recently sat in the same exams their students are preparing for. They know how teachers phrase questions, what gets marked and what doesn’t, where the tricky parts are. That institutional knowledge, recent, specific, human, is something no model has yet.

And then there’s the connection, students come back to the same tutors, and they know each other. They feel safe enough to ask the question they were too embarrassed to ask in class.

“That sense of belonging”, as Anna puts it, “is something AI simply cannot create.”

Lannee describes the qualities that make a great tutor: adaptive, flexible, curious, tech-savvy, creative. Someone who can pick up an unfamiliar topic, prepare it properly, and deliver it in a way that actually lands.

“Quality sessions,” she says, “come from tutors who genuinely care about the subject, about the student, and about getting it right.”

That care shows up in small ways: a tutor who stays patient when a student doesn’t understand, one who looks up the answer when they’re not sure rather than guessing, one who remembers that on the other side of the screen is a kid who’s stressed about their exam tomorrow.

Every lesson a student opens on StudyGo has been planned, prepared, and delivered by a person who took the time to understand what that student needs.

AI helps tutors move faster, handle more, and reach more students. But the judgment, the empathy, and the responsibility for what actually happens in that lesson always stay human.