If your child needs extra help, you now have three broad options: a human tutor, a free learning platform such as Khan Academy, or an AI tutor. Each has real strengths, and each has a way of letting families down. The right choice depends on your child, your budget and what is actually going wrong, and it is often a combination instead of a single pick.
Disclosure: we make Tutor AI. We have tried to describe each option fairly.
The short answer for most families is that a person is best for diagnosing a problem and keeping a child motivated, a free platform is best for core content and practice, and an AI tutor is best for patient daily practice and questions. The details below explain why, and where the evidence is thinner than the marketing suggests.
How do the three options compare at a glance?
| Human tutor | Khan Academy (free) | AI tutor | |
|---|---|---|---|
| Cost | Highest per hour | Free | Usually a subscription, lower than a human tutor |
| Availability | Scheduled sessions | Anytime | Anytime |
| Personalisation | High, from a relationship | Moderate (guided practice, mastery paths) | Can be high in the moment |
| Reading the child | Excellent | None | Limited |
| Motivation and accountability | Strong | Depends on the child | Depends on design and the child |
| Patience for repetition | Good, but time costs money | Unlimited | Unlimited |
| Reliability of content | Depends on the tutor | Strong, curated | Varies; check |
What do human tutors do well, and where do they fall short?
A good human tutor notices things no program can. They see frustration, boredom, or the "yup, I get it" that a child says when they do not, and they build a relationship that motivates a child to keep going. They can adapt on the fly, explain the same idea in several ways, and, with the right training, support specific needs. For a child who is stuck and losing confidence, that human connection is often what matters most.
The limits are cost, scheduling and consistency. Quality varies widely, not every area has good tutors, and one hour a week is not much practice on its own. The market has also been unstable. Varsity Tutors for Schools, one of the largest providers, shut its school service in 2026, and FEV Tutor ended abruptly in early 2025. Tutoring researchers quoted by The 74 said the providers in trouble were those that grew fast before building evidence of student gains. Ask any provider, human or AI, for evidence.
Is Khan Academy enough on its own?
Community discussions often describe Khan Academy as "free and explains things well," and that reputation is deserved for many topics. It offers video lessons, practice and progress tracking at no cost, with curated and consistent content, and students can work at their own pace with feedback on their answers.
The main limit is that it depends on the child's own motivation, and access does not equal use. Khan Academy has said only about 15 percent of students with access to its AI assistant regularly use it, and a randomised trial in Tennessee middle schools found that students using Khan Academy did modestly better than those in regular remedial maths, but the AI assistant added nothing to that gain (Hechinger Report). It is also less responsive to an individual child's confusion than a live tutor, and it is not tailored to your school's materials.
What can an AI tutor do, and what can't it?
An AI tutor is available at any hour and never tires of repetition. It can respond to the specific mistake a child has just made, it usually costs less than a human tutor, and the good ones ask questions and give hints, which fits what research suggests supports learning. For a child who needs a lot of patient practice, that combination is hard to match.
The limits are equally real. A tutor can be inaccurate, and general chatbots tend to give answers too quickly. It cannot read a child's face or mood, quality varies enormously between products, and some children will not engage without a person nudging them. The evidence is still developing. In a randomised trial with nearly 1,000 high school maths students (PNAS), an unrestricted AI helper reduced later exam performance by 17%, while a version designed to give hints avoided the harm without improving results. Design matters, and so does whether your child will use it.
What does each option cost besides money?
Price is only part of the picture, because each option also costs time and attention. A human tutor costs scheduling, travel or a video call, plus your effort to find someone good, and you may have to try more than one before it works. A free platform costs your supervision, since someone has to notice if a child is quietly skipping the practice or repeating only the easy exercises. An AI tutor costs a setup period and some checking, because you need to look at what it says, especially in maths, and watch that your child is thinking rather than copying. A cheap option that your child does not use is more expensive than a costly one that works.
What should you ask a human tutor?
If you go the human route, a few questions separate good tutors from average ones. Ask how they will find out what your child already knows in the first session, and what they will do when your child says "I get it" but does not. Ask how they will tell you what they worked on each week, and how you will both know after a month whether it is helping. A good tutor will answer without difficulty, and a vague answer to the last question is worth noticing.
When is a tutor not the answer?
Extra tutoring is not always the right response. If your child is falling behind across several subjects, is upset about school, or shows signs that suggest a learning difficulty, start with the teacher and, if needed, a specialist. Tutoring of any kind works best on a well-understood problem, and finding out what the problem is may matter more than choosing a tool.
How can you test what works for your child?
Rather than choosing once and forever, run a one-month experiment. Write down the single thing you want to change, such as "she stops avoiding fractions homework." Pick one option and use it for four weeks at a regular time, then ask the teacher what they have noticed and compare with where you started. Changing one thing at a time is the only way to know what helped, and a month is long enough to see a difference and short enough to change course.
Which option fits which situation?
If your child is capable but needs practice, a free platform plus an AI tutor for questions is a sensible pairing. If your child is stuck and losing confidence, a human tutor or teacher is the better start, ideally with AI practice between sessions. For a specific learning difficulty, get specialist support first and use AI only as a supplement. On a tight budget, free resources plus a well-designed AI tutor stretch furthest, and for a child who avoids asking for help, choose a tutor that starts the conversation, whether human or AI.
Many families do best by combining all three: a teacher or human tutor for diagnosis and motivation, a free platform for core content, and an AI tutor for daily practice and patient questions.
Where Tutor AI fits
Tutor AI is designed for the practice-and-questions part: it asks before it tells, uses interactive exercises, and goes back to what your child missed. It is not a replacement for a teacher or a caring adult, and we would suggest using it alongside the resources above.
Frequently asked questions
Is an AI tutor cheaper than a human tutor? Usually, but the cheaper option is only a saving if your child uses it. Compare cost per week of actual use, not the price on the page.
Can an AI tutor replace a teacher? No. The best evidence so far suggests that AI works as a supplement with careful design, and one small UK trial paired the AI with human tutors who reviewed its replies.
How do I know if it is working? Ask your child to explain something they practised, and ask the teacher what they notice. Improvement in class matters more than scores inside the app.
What if my child refuses to use it? That is common, and it is information. Try a different time or format, and consider whether a person, not a program, is what they need.