A day in the life¶
To make the product fully tangible, let's walk through a few real-life scenarios — how different people actually use ControlAI. The names and numbers are made up.
An owner's morning¶
08:40. Jamshid, the owner of a center, opens the admin panel on his phone over coffee. No need to call anyone or rewatch anything — the dashboard already shows the whole picture from yesterday.
At the top, in the "Needs attention" feed, there are three lines. The first is red: "Jasur Kh. — rating dropping: student talk-time 41%→28% over 3 weeks". Jamshid expands it — the cause is clear: the teacher is talking more and more himself (his share of talk has grown to 89%) and has almost stopped asking questions (down from 19 to 8 per lesson).
He doesn't guess and doesn't summon Jasur "to the carpet" blindly — he assigns the manager a task: schedule coaching on talk-time balance. Along the way he notices a green line: "Kamol T. — best week" — and decides to hold up his lessons as an example. The whole review took three minutes.
An academic manager's week¶
Tuesday. Malika, the academic manager (methodist), opens the method-compliance section. ControlAI has already filled in the rubrics for all of yesterday's lessons — she doesn't need to rewatch the recordings.
She works down the list. The AI marked most criteria confidently, with quotes from the lesson — those she simply confirms. A couple of items are flagged "needs a human look" (disputable, video-dependent) — on those she focuses her attention: she opens the moment of the lesson, watches, and enters her own verdict.
In one lesson the AI underrated "clarity of instructions" — Malika corrects it. That correction won't be lost: next time the system will score a similar case more accurately. In an hour she went through all of yesterday's lessons — previously that took a week of selective viewing.
A teacher's month¶
End of the month. Kamol opens his personal cabinet. Previously he didn't know in numbers how his lessons were going — now he sees: rating 4.7, up by 0.2; students speak 52%, 28 questions per lesson.
His goal is to bring student talk-time up to 55%. The progress bar is almost full. A tip suggests a bit more pair work in the second half of the lesson — Kamol tries that in his next lesson. A week later he sees in his own trend that the metric has gone up. For him this isn't "control from above" but a clear game of improvement — and he knows that a good rating — since the center has tied the rating to a bonus — also raises his income.
A branch manager's day¶
09:00. Malohat, a branch manager, starts her day in the "Operations" section. The list has two tasks: "Classroom No. 5 — camera offline" and "Classroom No. 4 — microphone needed". The first she hands to a technician for an audit; for the second she orders an inexpensive microphone.
Next she glances at the coverage summary: of 12 scheduled lessons, 11 were recorded, 9 have already been processed on time — the metric is healthy. One gap — a no-show lesson in Kids K-02 (the group didn't come). She notes it and moves on. By lunch tomorrow's schedule is checked, with no conflicts.
Key takeaway from this chapter: in real life ControlAI saves time and removes the "blind spots": the owner sees in three minutes where the problem is and why; the academic manager covers all the lessons instead of a couple; the teacher grows toward clear goals; the manager keeps operations under control.
→ Next: 12. What the center needs to provide — what is required to make all of this work.