Learning toys for a three-year-old, skill by skill

Walk down any toy aisle in India and roughly half the boxes say educational. It is the least useful word on the shelf, because it describes the packaging rather than the toy. A shape sorter is educational. So is a tablet loaded with phonics videos. They are not remotely the same purchase.

A more useful question is: what is my three-year-old actually working on right now, and does this toy get in the way of it or out of the way of it. Here is the honest version of that list.

Language, which is the big one

Between three and four a child's vocabulary roughly doubles, and the engine driving it is not exposure to words. It is turns. Researchers at MIT found that the number of conversational turns a child takes predicts language-related brain activity better than the sheer number of words they hear, and that the effect held regardless of family income or parental education. Turns, not volume.

So the test for a language toy is whether it creates a turn. A book you read together creates dozens. A puppet creates as many as you have patience for. A toy that plays a recorded sentence when a button is pressed creates none, however many words that sentence contains.

Turn-taking itself, which is a skill

Waiting, listening, and going second are not personality traits at three. They are skills, and they are practised. Simple board games with two rules, card matching, catch, and anything with an obvious go-then-wait rhythm all serve this directly. The Harvard Center on the Developing Child calls the underlying loop serve and return, and it is the same loop whether the returning is done by a parent, a sibling or a game.

Pretend play, which does more than it looks like

A cardboard kitchen, a doctor set, a box of mismatched figures. Pretend play is where a child rehearses other people's points of view, which is the foundation of both empathy and narrative. It is also the category most damaged by over-specification. A toy kitchen with twelve licensed characters and a sound chip tells the child what the story is. A wooden crate lets them decide.

Hands, which need less help than the boxes suggest

Threading, stacking, playdough, chunky crayons, tongs and a bowl of pompoms. Fine motor work at three is well served by cheap, dull-looking objects, and almost never improved by a screen. This is the part of the list where the ₹200 option is genuinely as good as the ₹2,000 one.

Numbers and letters, which are smaller than the marketing

Counting real things beats counting pictures of things, and both beat an app that counts for them. Three-year-olds benefit from number sense, meaning more and fewer and one-to-one matching, far more than from reciting to twenty. Almost any toy with countable parts does this. Very few toys sold specifically for it do it better.

What about screens

The published guidance is consistent and it is worth knowing before you shop. The World Health Organization recommends no more than one hour of sedentary screen time a day for children aged three and four, and says less is better. The American Academy of Pediatrics lands in a similar place for the years before five, with the strong caveat that it should be good content watched alongside an adult. The Indian Academy of Pediatrics published its own guidance in 2022, adapted for Indian households.

None of those bodies says a screen is poison. What they consistently point at is the cost of what the screen replaces, which at three is almost always conversation and movement.

A short buying rule

Prefer toys that wait for your child to act over toys that perform at them. Prefer open ends over scripts. Prefer anything that produces a turn. And be suspicious of any box where the word educational is doing work that the toy inside is not.

That rule is the whole reason Kheelu is built the way it is: it answers, then asks, and it has no screen anywhere. If you want the same rule applied to every AI toy on the market rather than just ours, the guide to choosing one sets out the five checks.