AI tutor
What an AI Tutor Actually Does
Most learners do not get stuck during the lesson. They get stuck at night, on question four, with nobody to ask. An AI tutor is the patient second explanation — grounded in what the institution teaches, and visible to the teacher afterwards.
Tutor session
Student, 10:40pm
I keep getting question 4 wrong. Can you just give me the answer?
Tutor
Not yet. Your setup is right and the substitution step is where it breaks. Redo that one line and I will check it with you.
Saved to the learning record
Simultaneous equations. Second session this week. Visible to the teacher tomorrow morning.
Illustrative interface. Not customer data.
What this actually means
A search result answers. A tutor explains.
An AI tutor is a conversation that stays on one difficulty until it is gone. It asks what the learner already tried, works out where the reasoning broke, and explains that step — again, differently, as many times as it takes, without sighing.
It works from your material
The tutor is pointed at the syllabus, the course content and the assessments the institution uses, so its explanation matches the one the class was given.
It holds a conversation
A learner can say they still do not understand. That is the whole point. The explanation changes approach rather than repeating itself word for word.
It leaves a trail
Sessions are saved against the learner. The teacher sees what the class struggled with; the learner gets their own explanations back as revision.
What it changes
Help arrives at the hour it is needed
-
A learner who is stuck at 11pm is not stuck until Monday.
The tutor is available outside the timetable, so a gap that would have compounded over a weekend gets closed the night it appears.
Always-available explanation
-
Ask the obvious question without asking it in front of the class.
Plenty of learners will not raise a hand. A private conversation removes the social cost of not understanding something everyone else seems to have understood.
Private practice
-
Learn in the language the learner thinks in.
Explanation can switch language while the subject vocabulary stays as the course uses it, which matters in classrooms where the medium of instruction is a second language.
Multilingual explanation
-
Practice aimed at the thing that broke.
The tutor can generate a few more questions on exactly the step the learner got wrong, rather than sending them back through the whole chapter.
Targeted practice
-
Teachers find out what the class did not follow.
When eleven students ask the tutor about the same step, that is a signal about the lesson, not about the eleven students. It reaches the teacher as a summary.
Session insight for educators
The limits
The interesting part is what it refuses to do
A tutor that hands over answers is a homework machine, and every teacher can tell within a fortnight. The constraints below are design decisions, not missing features.
- It works the learner towards the answer rather than supplying it
- It stays inside the institution's material instead of the open internet
- It does not grade formal assessments or decide what a learner is worth
- It does not set the curriculum; the teacher or faculty member still does
- Sessions are visible to the educator responsible for that learner
Where the teacher stays in charge
- The material
- Loaded and approved by the institution, so the tutor teaches the syllabus rather than a version of it.
- The record
- Every session belongs to the learner's file. Nothing about the tutor is hidden from the teacher.
- The judgement
- What a learner has mastered is confirmed by assessment and by a person, not asserted by a chat window.
Questions
What learners and teachers ask
Will students use it to cheat on homework?
The tutor is built to walk a learner through a problem rather than complete it, and it will say no to a request for a finished answer. It is not a perfect defence against a determined student, which is why sessions are recorded against the learner and visible to the teacher. A pattern of asking for answers is something the teacher can see and address.
Does the tutor know what my class is being taught?
Yes, when the institution loads its material. The tutor is grounded in the syllabus, course content and question banks the school or college uses, so its explanations line up with the lesson instead of contradicting it. Where it has nothing relevant to draw on, it says so rather than inventing something.
Is it for school students or college students?
Both, with different vocabulary and different depth. In schools it works at concept level against the class syllabus. In higher education it works at course and competency level, and it can be pointed at the assessments a programme uses. Enterprise learners get the same conversational help against role skills.
What happens to the conversations a learner has?
They are stored inside that institution's own tenant, attached to the learner's record, and available to the educator responsible for them. They are not pooled across customers. A learner can revisit their own sessions, which is usually the most honest revision material they have.
See the tutor answer a question from your own syllabus
Bring a chapter, a course module or a past paper. We would rather show you the tutor working on material you recognise than run a scripted tour.
Keep reading
-
AI
The AI layer underneath the tutor, the assistant and the analytics.
-
Learning paths
The route the tutor sits inside when a learner is working through a gap.
-
Schools
The tutor in a school day, alongside teacher and parent views.
-
Higher Education
The same tutor at course and competency level for colleges.
-
For students
What a learner actually sees when they log in.
-
The platform
Assessment, learning and analytics in one place.