Every tutoring advertisement you have seen in the last two years has promised personalized learning. Almost none of them tell you what they mean by it, and that is not an accident. The phrase is doing an enormous amount of work while committing to almost nothing.
This matters because you cannot evaluate a claim you cannot define. If a platform says it personalizes your child's learning, the useful question is not whether that is true but which kind of personalization it is offering, and whether that kind is worth paying for.
The word is doing too much work
In practice, "personalized learning" describes three genuinely different things. Products blur them together because blurring flatters the weakest one. Separating them is most of the work of becoming a competent buyer.
1. Pace: when you move on
The oldest and simplest form. Every student works through the same material in the same order, but nobody advances until they have demonstrated they can do the current thing. A fast student finishes a chapter in two days; a slower one takes eight, and is not dragged along behind the class.
This is the easiest kind to build and the easiest to verify. If a system will not let a student move to the next topic until a check is passed, it personalizes pace.
2. Path: what order you meet topics in
Harder. Here the system changes the sequence, not just the speed. A student who is fluent with fractions but shaky on place value is routed backwards to place value first, because one is built on the other. Two students in the same class can be working on genuinely different topics on the same afternoon.
Path personalization requires the system to hold a model of how topics depend on each other. Some products have this. Many claim it while actually doing something closer to pace.
3. Difficulty: how hard the next question is
The narrowest kind, and the one most often sold as the whole thing. The content and the order stay fixed; only the difficulty of individual questions moves up or down based on recent answers.
This is real and useful. It is also what most "adaptive" quiz apps mean, and it is considerably less than what "personalized learning" sounds like it promises.
What a system needs before it can adapt anything
All three kinds depend on the same precondition: the system must have some model of what this particular student currently knows. Without that, adaptation is guesswork dressed up as intelligence.
That model has to come from somewhere, and there are only a few honest sources: questions the student has answered, work the student has submitted, and time spent struggling. A tool that asks nothing and observes nothing cannot personalize anything, no matter what the marketing page says. This is the single most useful test you can apply, and we go into it further in AI study assistant vs AI tutor.
Where the evidence is strong, and where it is not
Personalization is not a new idea, but generative AI has changed what is buildable, and the research has started to catch up. Two studies from 2025 are worth knowing about, because they point in opposite directions and both are solid.
In a randomised controlled trial at Harvard, published in Scientific Reports, 194 undergraduate physics students were assigned either to a purpose-built AI tutor or to an active-learning classroom taught by experienced instructors. The AI-tutored group learned more, and did so in less time.
In a separate study of roughly a thousand high school students published in PNAS, students given unrestricted access to GPT-4 improved sharply on practice problems. When the tool was taken away for the real exam, they performed worse than students who had never had access at all.
The difference between those two results is not the technology. It is the design. One system was built to make students think; the other was available to think for them. We unpack both studies in Is AI good for students?.
Personalized does not mean automated
The most common misreading of personalized learning is that it means learning without a teacher. Nothing in the evidence supports that, and the PNAS result argues directly against it.
Personalization changes what a student is asked to do next. It does not, on its own, supply the reasons a twelve-year-old keeps going on a Wednesday evening when the work is hard and boring. It does not notice that a child has gone quiet for three sessions. It does not decide that this week, the right move is to stop doing maths and rebuild some confidence.
Those are judgement calls, and they are the part of teaching that has survived every previous wave of educational technology. The useful arrangement is not AI instead of a teacher but AI alongside one, which we describe in AI plus a human teacher.
How many students are already doing this anyway
Whatever any of us conclude, students have largely made their own decision. Pew Research Center found that the share of US teens using ChatGPT for schoolwork doubled from 13% in 2023 to 26% in 2024. In more recent fieldwork, 54% reported using chatbots for help with schoolwork.
That is not a personalized learning system. It is unsupervised access to a general-purpose tool, which is the exact condition the PNAS study found to be harmful. The realistic question for most families is not whether AI enters their child's studying but whether it does so with any structure around it.
Questions worth asking any platform
Before you pay for personalized learning, ask which of these the product actually does:
- Does it stop a student advancing before they are ready, or only suggest?
- Can two students end up on different topics, or only at different speeds?
- What does it observe about my child in order to adapt? If the answer is nothing, nothing is adapting.
- Does it ever make the work harder on purpose, or does it optimise for the student feeling successful?
- When my child is stuck, does it explain, or does it supply the answer?
That last one matters more than the rest combined, and it is the difference between the two research results above. There is a longer version of this in How to choose an AI tutor.
Our position, and the part that already works
We should be straightforward about this, since we are a tutoring company writing about tutoring technology. eGunas today is one-to-one teaching by human tutors. We are building AI-assisted learning and it is not live yet, so nothing above is a description of a product we are currently selling.
The clearest and least speculative version of adapting to a student is accessibility: removing a barrier so a child can reach work they already understand. We have covered what assistive technology does for dyslexia and ADHD separately, because it is the one area where the evidence is settled rather than promised.
We wrote this because the vocabulary is genuinely confusing and most of the explanations available are written by people with something to ship. If it helps you ask a competitor a harder question, that is a reasonable outcome.
How eGunas helps your child
eGunas teaches one child at a time. A session follows the student's own board syllabus but is built around what that particular student needs next -- their pace, the topics they are actually weak at -- rather than one lesson delivered to thirty students at once. That is what personalised learning means in practice.
eGunas AI, our AI-powered learning platform, is planned for 2026 and not available yet. It is being designed to extend that same individual attention beyond the lesson hour: practice pitched at the right difficulty, available whenever your child sits down to work, so progress can continue between sessions instead of waiting for the next one.
- Browse our courses -- CBSE, ICSE and state boards, Classes 1 to 10
- How eGunas online tutoring works -- one-to-one lessons online, wherever in India you live
- Book a free demo class -- one-to-one with a tutor, at no cost
- Ask us about your child -- tell us the subject and year, and we will say honestly whether we can help
New articles are published on the eGunas blog.
