What Does a Good Private Tutor Actually Do That a Textbook or AI Cannot?
Hook: Sometimes the important question isn't “What don't you know?” but “Why do you think that?”
We have never had more ways to learn.
A student struggling with quadratic equations, photosynthesis, electric circuits or chemical bonding can open a textbook, watch a YouTube lesson, search the internet or ask an AI system for an explanation in seconds.
That is enormously valuable.
In fact, I think good tutors should encourage students to use these resources rather than pretend that private tuition is the only effective way to learn.
But it raises an interesting question.
If a textbook can explain the topic, YouTube can demonstrate it and AI can answer questions almost instantly, what exactly is a good private tutor being paid to do?
The answer, I would suggest, is not simply to explain things.
It is to diagnose learning.
A good tutor is continually trying to discover not merely what answer a student has given, but why they gave it.
And sometimes that small distinction changes the whole lesson.
Knowing That an Answer Is Wrong Is the Easy Part
Imagine that a student is asked:
Solve: 3(x + 2) = 15
They write:
3x + 2 = 15
and eventually produce the wrong answer.
It would be very easy simply to correct them:
3(x + 2) = 15
3x + 6 = 15
3x = 9
x = 3
Problem solved?
Possibly not.
Because the interesting part is not that the student obtained the wrong answer.
The interesting part is:
Why did they write 3x + 2?
Perhaps they have never properly understood the distributive law.
Perhaps they know it but forgot to apply it.
Perhaps brackets make them nervous and they rushed.
Perhaps they think multiplication applies only to the first term.
Perhaps they can expand 3(x + 2) but become confused when the number outside the bracket is negative.
Those are very different problems.
Giving the student another twenty identical questions might help one of them and do almost nothing for another.
That is where diagnosis begins.
The Most Useful Question Can Be: “Why Did You Do That?”
One of the most powerful things a tutor can say is remarkably simple:
“Talk me through what you were thinking.”
That question can reveal far more than another page of exercises.
Consider a GCSE science student who says:
“Increasing the voltage increases the current because the electricity gets pushed through faster.”
That answer contains something useful.
The student has recognised that voltage and current are related.
But the word “faster” may reveal a deeper misconception.
Do they believe current is the speed of the electrons?
Do they understand charge flow?
Do they understand resistance?
Do they know what potential difference actually represents?
Instead of immediately saying, “That's wrong,” I might ask:
“What do you mean by faster?”
Then:
“What exactly is moving?”
And perhaps:
“If twice as much charge passed a point every second, what would happen to the current?”
Now the lesson has become diagnostic.
We are finding the structure of the student's understanding.
A Textbook Cannot See You Hesitate
A textbook can contain an excellent explanation.
What it cannot do is watch a student read that explanation.
An experienced teacher notices surprisingly small things.
A student may pause for five seconds before answering a question they supposedly understand.
They may confidently answer one type of question but immediately reach for their calculator when the numbers are presented differently.
They may repeatedly look back at a formula sheet.
They may whisper a remembered rule to themselves.
They may erase the same line three times.
They may answer correctly but look completely unconvinced.
Or they may answer incorrectly with enormous confidence.
All of those behaviours provide information.
Sometimes I am more interested in the hesitation before an answer than the answer itself.
Because hesitation can indicate that knowledge has not yet become secure.
Getting the Right Answer Can Hide a Problem
This is one of the traps of teaching.
A student produces the correct answer.
We congratulate them.
We move on.
But occasionally they arrived at the answer for completely the wrong reason.
Suppose a student is asked about acceleration and correctly selects:
a = (v - u) / t
Excellent.
But then I might ask:
“Why did you choose that equation?”
If the answer is:
“Because it was the only formula with all those letters in it.”
we have discovered something important.
The student has not necessarily understood acceleration at all.
They have learned a formula-selection strategy.
That strategy might work on straightforward examination questions.
It becomes much less reliable when a question requires several stages of reasoning.
So rather than giving another nearly identical question, I might deliberately change the situation.
I might ask about a car slowing down.
Then an object travelling in the opposite direction.
Then a graph.
Then I might remove the equation entirely and ask the student to describe what acceleration means.
The next question is being selected because of the previous answer.
That is tutoring at its most useful.
Good Tutoring Is Not a Script
There is another important difference between teaching material and teaching a person.
A worksheet has an order.
Question 1 leads to Question 2, then Question 3.
A good tutoring session does not always work like that.
I might prepare twelve questions and only use five of them.
Why?
Because Question 3 might reveal something unexpected.
Perhaps the student cannot manipulate fractions.
That changes everything.
There is little point continuing with increasingly difficult algebra if the real obstacle is arithmetic from several years earlier.
So we temporarily leave the planned lesson.
We investigate fractions.
Perhaps we discover that adding them is fine but division by a fraction is not.
We deal with that.
Then return to the original problem.
A successful lesson may therefore look quite different from the lesson I originally planned.
That does not mean the planning was wasted.
It means the student's thinking has determined what happens next.
The Next Question Matters
This is perhaps the part of tutoring that is easiest to underestimate.
Once I know what a student can do, I need to decide:
What should I ask next?
Too easy and I learn very little.
Too difficult and I may simply demonstrate something the student cannot yet do.
The most revealing question is often one small step beyond what the student has just demonstrated.
For example, a student successfully expands:
4(x + 3)
So rather than giving:
5(x + 2)
I might try:
4(x - 3)
If that works:
-4(x + 3)
Then:
-4(x - 3)
Each question introduces a small additional complication.
Eventually we find the boundary between what is secure and what is not.
That boundary is extraordinarily useful.
It tells us where teaching needs to begin.
Sometimes the Problem Is Not the Subject
A student can apparently struggle with mathematics when the underlying difficulty is reading.
They can struggle with physics when the real problem is rearranging equations.
They can struggle with chemistry because they have never become confident with ratios.
They can struggle with biology because they understand the topic but cannot convert their knowledge into a six-mark examination answer.
This is why saying:
“They're struggling with electricity.”
may not be precise enough.
What exactly are they struggling with?
The concepts?
The mathematics?
Circuit diagrams?
Formulae?
Units?
Interpreting graphs?
Reading the question?
Knowing which information is relevant?
Writing an explanation?
A tutor should gradually narrow the problem down.
The Difference Between Knowledge and Performance
Another important part of diagnosis is recognising that a student may know something without being able to demonstrate it effectively.
I sometimes find that a student can explain a scientific idea perfectly well in conversation.
Ask them:
“Why does sweating cool the body?”
and they can give a perfectly sensible verbal explanation.
Put essentially the same question on an examination paper and they write two vague sentences and score one mark.
Their scientific knowledge may not be the problem.
The missing skill could be examination technique.
They may not understand command words.
They may not know how much detail is expected.
They may not connect their answer explicitly enough to the question.
That requires a different intervention from simply teaching evaporation again.
Confidence Can Be Misleading Too
There are students who constantly say:
“I can't do this.”
when actually they can.
And there are students who say:
“Yes, I understand.”
when they absolutely do not.
A good tutor gradually learns not to rely entirely on either statement.
Instead, we test understanding gently.
“Show me.”
That is often much more revealing.
If a student tells me they understand factorisation, I might give them a fairly straightforward example.
Then one that looks slightly unfamiliar.
Then perhaps ask them to create their own example.
Being able to explain something or construct an example is often stronger evidence of understanding than simply recognising an answer.
But Doesn't AI Do Some of This?
Increasingly, yes.
And I think it would be rather foolish for tutors to pretend otherwise.
Modern AI tools can be extremely useful educational resources.
A student can ask:
- “Explain this more simply.”
- “Give me another example.”
- “Don't tell me the answer; give me a hint.”
- “Make me five questions that get progressively harder.”
- “Explain where I went wrong.”
- “Test me on this topic.”
Those are excellent uses of AI.
AI can also patiently explain the same concept several different ways without becoming frustrated.
That makes it a potentially powerful companion for independent study.
So the argument should not be:
Human tutor good. AI bad.
That misses the point.
The more interesting question is:
What information does each have available when deciding what the learner needs next?
AI Can Respond to What the Student Shows It
An AI system usually works from the information the student provides.
And that creates an interesting problem.
Students are not always good at diagnosing their own misunderstandings.
A learner might ask:
“Can you explain electromagnetic induction again?”
But perhaps electromagnetic induction is not really the problem.
Perhaps they do not understand magnetic fields.
Or perhaps they understand the physics but cannot interpret the diagram.
Or perhaps the mathematics is causing the difficulty.
An experienced tutor can sometimes realise:
We need to go backwards before we can go forwards.
And that may involve teaching something the student did not even realise they needed.
Sometimes I Deliberately Let a Student Be Wrong
This can feel strange.
A student starts solving something incorrectly.
The temptation is to stop them immediately.
But sometimes I don't.
I let them continue for another line or two.
Why?
Because I want to find out where their reasoning leads.
Correct them too quickly and I may only discover the first visible mistake.
Let them explain their reasoning and I may discover the misconception underneath it.
Of course, I do not want them practising an incorrect method repeatedly.
But a few carefully observed seconds of wrong reasoning can tell me far more than instantly supplying the correct method.
One Wrong Answer Can Lead to Five Different Lessons
Imagine five students all incorrectly calculate the gradient of a graph.
The same wrong answer could arise because:
- One student does not understand what gradient represents.
- One swaps change in y and change in x.
- One chooses inappropriate points.
- One makes an arithmetic mistake involving negative numbers.
- One understands everything but copies a coordinate incorrectly.
Should all five students receive exactly the same explanation?
Probably not.
And this is why simply recording:
Question 6 – incorrect
does not tell us enough.
A tutor needs to ask what kind of incorrect answer it was.
The Tutor Is Also Looking for Patterns
Individual mistakes matter.
Patterns matter more.
If a student makes one sign error, it may simply be carelessness.
If they make six sign errors in twenty minutes, something else may be happening.
If they repeatedly lose marks on questions containing graphs, I become interested in graphs.
If every calculation is correct until units are introduced, I become interested in units.
If they perform well until questions contain several paragraphs of information, perhaps we need to work on extracting relevant data.
Good tutoring therefore involves building up a picture over time.
Not:
“Can this student answer this question?”
but:
“What sort of questions does this student find difficult, and what do those questions have in common?”
A Good Tutor Should Eventually Make Themselves Less Necessary
There is another important point.
The aim of tuition should not be to create dependency.
If the student can only solve a problem while the tutor is sitting beside them saying:
“What do you do next?”
then we have not finished the job.
Support should gradually reduce.
Perhaps the progression looks something like this:
Tutor demonstrates -> student works with prompts -> student works independently -> student explains the method -> student applies it to an unfamiliar problem
That final stage is important.
Examinations rarely reward students simply for reproducing the exact example they saw in a lesson.
They need enough understanding to recognise the underlying idea when it appears in a different form.
Books, Videos, AI and Tutors Are Not Competitors
I think education becomes much more interesting when we stop asking which resource is “best”.
They do different jobs.
A textbook provides structured coverage.
A good video can provide an excellent visual explanation.
An AI system can provide instant interaction, practice and alternative explanations.
A teacher provides curriculum structure and the experience of learning alongside others.
A tutor provides something particularly personal:
continuous diagnosis of one learner's thinking.
The strongest learning environment may use all of them.
A student could watch an explanation before tuition, work through questions with their tutor, use AI afterwards for additional practice and then return with the questions they still cannot solve.
That is considerably more powerful than pretending one method must replace all the others.
Perhaps the Tutor's Most Important Tool Is Curiosity
Good tutors need subject knowledge.
They need to explain clearly.
They need suitable questions.
But I increasingly think another quality matters just as much:
curiosity about why the student thinks what they think.
Why did you choose that equation?
Why did you draw that diagram?
Why did you cross that answer out?
Why were you confident about Question 3 but not Question 4?
Why does this method make sense to you?
What would happen if I changed this number?
Can you explain it without using the formula?
Those questions turn tuition from answer-correction into diagnosis.
And once we understand the misconception, we can actually do something about it.
The Real Value of a Good Tutor
It is easy to describe private tuition as additional teaching.
Sometimes it is.
But at its best, I think it is something rather more interesting.
A good tutor is constantly gathering evidence.
Watching.
Listening.
Questioning.
Testing.
Changing the explanation.
Selecting the next problem.
Going backwards when necessary.
Moving forwards when the student is ready.
And gradually handing responsibility back to the learner.
A textbook may contain the explanation the student needs.
A video may demonstrate it beautifully.
AI may explain it in ten different ways.
But sometimes the most valuable moment in a lesson occurs when a student confidently gives the wrong answer and the tutor does not immediately correct them.
Instead, they ask:
“That's interesting. Why do you think that?”
Because once we understand why a student is wrong, we are much closer to helping them become right.
Perhaps the real skill of private tutoring isn't knowing all the answers. It is knowing which question to ask next.
