Admin panel for the quality-control office (the "academic manager" role)¶
Who this is. "Academic manager" is the product name for the role; at the university it's held by a staff member of the quality-control office (ichki ta'lim sifati nazorati bo'limi) — or, at the university's discretion, by the head of department acting as reviewer for their own subjects. This is the role the human reviewer turns into: previously an open lesson (ochiq dars) was held once or twice a year, with a committee sitting in the classroom filling out the observation form by hand. Now ControlAI fills out this form for every class session, and the expert reviews and corrects the AI's ready-made assessments. Their corrections aren't wasted — they make the system more accurate over time (our "accumulation effect").
In the current MVP there is no separate "academic manager" role yet — its functions are handled by the general "manager" role. What follows is the target picture.
What they see and can do¶
Reviewing the observation form (main screen)¶
For every class session, ControlAI fills out the university's own observation form — typically 12–18 criteria. The expert opens the completed form and sees, for each criterion: the AI's score, a short quote as evidence from the class, and how confident the AI is. Their job is to confirm or correct the verdicts. The AI itself flags disputed and subjective items as "needs a human look" — those are what the expert checks first.
Every correction the expert makes is saved and sharpens future assessments — so over time the AI makes fewer mistakes, and the office is left with less and less manual work. Oversight ends up covering every class session of the semester, not one or two open lessons a year.
Topic coverage¶
By group, you can see what has been covered from the course syllabus, what's partially covered, and what hasn't been covered yet — accumulated across the semester. Useful both for oversight and for planning.
Configuring the methodology (university profile)¶
The quality-control office (together with the rector's office) sets the "rules of the game" for the AI:
- uploads the lesson observation form — the university's own template;
- uploads the syllabi and textbooks (what gets covered, and in what order, during the semester);
- configures review emphasis to match the university's requirements — including what's specific to lectures, seminars, and practical classes.
Class sessions, transcripts, analytics — for context¶
The expert sees the feed of class sessions, transcripts, metrics, and instructor analytics — to judge quality against the full picture, not a single criterion. Consents are view-only.
What this role typically doesn't do¶
- Doesn't handle billing or workload accounting — that's the rector's office.
- Doesn't manage equipment or the schedule at the configuration level — that's the dean's office (though it does see the results).
Bottom line: the quality-control office is the "human expert" layered on top of the AI. ControlAI fills out the observation form for every class session, while the expert reviews, corrects, and sets the university's methodology. This lets the university extend oversight to every class session instead of one or two open lessons a year — while still keeping genuine human expertise in the loop.
→ Next: Head of Department · back to overview