What Parents Expect From a Tutor Now That They Have AI at Home

By Brandon Collis 3 August 2026 10 min read

ai-tutoringpersonalised-tutoringtutoring-businessclassquillb2b

What Parents Expect From a Tutor Now That They Have AI at Home

What do parents expect from a tutor in the age of AI?

Parents expect three things a free chatbot cannot give: work tailored to their child’s specific gaps rather than a generic worksheet, visible evidence that it is actually working, and a coach who reads the student in real time and knows what to do next. The tutor’s job was never just answering questions. It is explaining, spotting a weakness as it happens, motivating a tired teenager and, honestly, a bit of life-coaching too. AI only goes so far. What has changed is that a tutor can now use AI to make every session more tailored, and parents paying a premium will increasingly expect to see that difference.

That shift is quiet, but it is already changing what wins and loses a family.

The bar is moving, and a generic worksheet will not clear it

For as long as tutoring has existed, part of what a parent paid for was access to someone who knew the material. That part is becoming free. Students are starting to have capable AI in their pockets that will explain quadratics, mark an essay, or generate ten practice questions on demand, at any hour. It is no longer hypothetical: Claude Code recently released a free year for students and teachers in the United States, and access like that only spreads. As it does, the ability to personalise learning keeps climbing, and so does the baseline a family measures you against.

No parent has sat me down and asked why they should pay me when they have a chatbot at home. But the expectation is coming, and it is worth getting ahead of. A parent paying a premium price will, reasonably, expect a premium level of service to match what their child can now get for nothing. If what your tutors hand over is a photocopied worksheet and an explanation a parent could have generated themselves, you are quietly competing against a free tool that does the commodity half of the job faster than you can.

This does not make tutors obsolete. It makes the generic parts of tutoring obsolete. The companies that win in this environment are not the ones that avoid AI or the ones that lean on it hardest. They are the ones that understand which half of tutoring became a commodity and double down on the half that did not.

Why a parent still pays for a tutor, in one line

Here is the honest answer to the question parents will eventually ask out loud. A tutor is a coach, an explainer and a real-time reader of a student’s weaknesses, and no chatbot does that. AI can hand a student a correct explanation. It cannot tell that the student has gone quiet because they are lost rather than bored, cannot decide that a “careless” mistake is actually a misunderstanding three topics back, cannot hold a teenager to account week after week, and cannot motivate a kid who has decided they are “just bad at maths”. That is the half of tutoring that got more valuable, not less, the moment information became free.

So the role did not shrink. It sharpened. The parts of a session that a machine can do well are exactly the parts a parent stops paying for, which throws the human parts into relief: the explaining, the judgement, the coaching. A tutor who leans into those is not threatened by a free chatbot. They are the reason a family keeps paying while the free chatbot sits unused on the kitchen table.

What parents actually want now: three shifts

1. Tailored, not generic. A parent whose child can get infinite generic questions for free expects the paid version to be aimed at their child, at the specific topics that child keeps getting wrong. “Here is a worksheet on integration” is a commodity. “Here is a set built from the three question types your child lost marks on last week” is not. The expectation has moved from volume to precision.

2. Visible evidence, not reassurance. A parent who can check anything instantly is far less satisfied by “he’s doing well” than parents were five years ago. They expect to see it: which topics have improved, what their child actually attempted, where the gaps still are. Vague reassurance now reads as a company that either is not tracking progress or does not want to show it.

3. The human judgement AI cannot give. This is the half described above, and it is what a parent is really buying. Deciding what a specific child needs next, reading frustration and easing off, holding a student to a standard. No chatbot does this. The parent knows it, even if they cannot articulate it, which is exactly why they still pay for a human.

AI just put big-company resources within reach of every tutor

There is a genuinely good side to all of this, and it is the part most “AI and tutoring” pieces miss. Comprehensive, high-quality resources used to be something only large tutoring companies could offer. A solo tutor or a small operation simply did not have the time to build a full bank of material, so they leaned on generalised worksheets: a topic, a few easy questions, some harder ones, an extension set, the same for every student. AI has changed the economics of that. The kind of resource library that was once a big-company advantage is now within reach of an independent tutor, and the work no longer has to be generalised per topic. It can be tailored to the specific gap in front of you.

A small example from my own tutoring. One student was struggling with inverse operations in algebra, undoing something like 3x or x/2, the idea that the opposite of division is multiplication and so on. Generic algebra worksheets were not the problem she needed solved. So I built her a step-by-step worked-example sheet aimed squarely at that one misunderstanding. That kind of targeted resource used to take real time to make, which is exactly why most tutors reached for the generic sheet instead. Being able to produce it quickly is what AI actually changes for a tutoring business.

See an example of that kind of targeted worksheet (PDF): the same undoing-operations idea laid out as worked examples with a short practice set.

The honest limit: why raw AI generation is not enough

Here is the part I will not pretend around, because getting it wrong burns trust with parents fast. Raw AI question generation is not good enough on its own. It often is not specific enough to a real curriculum, and it sometimes gets the maths wrong. Handing students AI-generated questions unchecked is a way to teach them something incorrect with total confidence, which is worse than a dull worksheet.

The model that actually works is a vetted, genuinely good question set that AI then tailors to the individual student. That is deliberately how ClassQuill is built, and it is the opposite of “AI spits out questions”. Underneath is a curriculum graph that understands each topic and what it involves, and a bank of base questions that have been checked. AI tailors on top of that foundation rather than inventing from nothing. The result is work aimed at a specific child’s gap that is still curriculum-correct, because the maths was right before the AI ever touched it.

How to meet the new expectation without adding tutor hours

Here is the trap: meeting a higher bar sounds like more work for your tutors, and tutor time is the one thing you cannot cheaply add. The resolution is to use AI for the commodity half so your tutors spend their time on the half that matters.

Tailoring homework by hand does not scale. Doing it as a by-product of work the student already does, does. When a student’s homework is auto-marked against that curriculum-aligned question bank, the result flags the specific step where it went wrong, and the next set can be built targeted at that gap rather than being the next page of a textbook. A tutor can accept it, tweak it, or build their own from the bank, but they are not starting from a blank page every week. The tutor also opens each session already knowing which topics that student is weak on and how that has shifted, instead of reconstructing it from memory. The AI does the sorting and the marking; the tutor makes every teaching call.

The change this made for me was concrete. I no longer waste the start of a session asking a student what they want to cover or re-diagnosing where they are, because I already know. Students can save the questions they got stuck on during the week, so nothing is forgotten by the next session. Classes became far more targeted, students got more out of them, and both student and parent satisfaction went up, which is what quietly drives referrals. That is the payoff of using AI for the sorting so the human keeps the teaching. (For the in-session version of this, see how tutors can use AI to enhance sessions; for the honest read on which AI claims are real, see what’s real and what’s hype.)

The part parents can see: making the tailoring visible

Tailored work that a parent cannot see is only half the value. The second shift, visible evidence, is what turns “we do personalised work” from a claim into something a parent can watch happen. When the same marked work that targets a child’s gaps also rolls up into a topic-by-topic progress view the parent can open, the parent stops needing to take your word for it. They can see the work was aimed at the right things, and that it moved.

That visibility is not a nice-to-have. It is the retention mechanism. Parents do not cancel because tutoring is expensive; they cancel when they cannot tell whether it is working (the full argument is in why tutoring clients really cancel). In the AI era, when a parent’s baseline expectation for visibility is higher than ever, progress a parent can actually see is the difference between a family that renews and one that quietly drifts. The learning platform that generates the tailored work and the progress view is the same system, which is the whole point: the tailoring and the proof come from one place, not two.

The line that matters: AI makes the work tailored, the tutor makes it matter

The tutoring companies that struggle over the next few years will be the ones that treat AI as either a threat to hide from or a replacement to hand the teaching to. Both misread what parents want. AI has not replaced the tutor; it is raising the bar for what a tutor is expected to deliver, and it has handed tutors the tool to clear it. The work gets more tailored and the evidence gets more visible, which is exactly what a parent whose child has free AI at home will come to expect, and the tutor is freed to spend their hours on the coaching and judgement no model can provide.

Get that division of labour right and “we use AI” stops being a nervous disclaimer and becomes the reason a parent chooses you: not because you have AI, but because your students get work aimed at their actual gaps, checked against a real curriculum, and their parents can see it working. That is what tutoring in the age of AI is supposed to look like, and it is what ClassQuill’s learning platform is built to do.

Related reading