A parent’s first 30 minutes in the studio.

Minute zero: opening the studio.

Your kid double-clicks the icon. The studio loads in a few seconds and asks for their name. Then it asks them to pick a track: Game, Story, Site, or Movie. These are the four worlds the studio knows how to make. Your kid will probably hover the cursor over all four and then pick Game, because they’re a kid and games are concrete. That’s fine. They can come back to the others later. The license covers all four.

If they ask you which one to pick, the right answer is “whichever one sounds most fun.” Resist the urge to steer. The first moment of agency in the studio is choosing the track, and you want them to feel that the choice was theirs. We’ve watched this play out fifty times in cohort intake. Kids whose parents pick the track for them spend the rest of the session looking to the parent for permission. Kids who pick their own track start moving.

After the track, the studio shows a small landing screen with a few starter prompts. “Make a platformer.” “Make a top-down adventure.” “Start from scratch.” Again: let the kid pick. If they go with a starter, the studio will scaffold a tiny working game. If they pick from scratch, the studio will open an empty project. Both are fine. Both produce something the kid can change in the next few minutes.

What you’re watching for in this first minute is the moment your kid stops looking at you and starts looking at the screen. That’s the goal. The studio is designed to earn the kid’s attention on its own. Your only job is not to break it.

Minute five: the first wizard.

The studio uses small, scoped helpers called wizards. Each wizard does one thing: “add a character,” “add an enemy,” “make this scene scarier.” They show up in a panel on the side. Your kid will pick one within the first few minutes, almost always “add a character.” They want to put themselves into the game.

The wizard opens a small form. It asks the kid for the character’s name. Sometimes a description. Sometimes a color. The form is doing real work: it’s making your kid specify a few details before the AI does anything. If they want a dragon, they have to say green dragon with red eyes or tiny rainbow dragon. The specificity is the lesson. The AI will fill in the rest, but the kid is naming the thing.

If your kid asks you what to name the character, this is one of the moments where it matters what you say. Don’t suggest a name. Ask them “what’s their personality?” or “what do they sound like?” The kid will answer, and the name will come out of their own answer. You’ve made it theirs without taking the move from them. This is the autonomy-support move Edward Deci and Richard Ryan write about in self-determination theory: you scaffold the choice without making it for them.¹

The kid presses Generate. The wizard hands the request to the AI. A few seconds later, the studio shows a card called the ChangeDisclosure card. The card says, in plain English, exactly what’s about to change. “A new character named Pip will be added to the level. They are a green dragon with red eyes.” The kid has to decide: Keep, Review, or Undo.

Minute seven: the first AI proposal.

This is the moment in the session that does the most pedagogical work. The kid is looking at a proposed change. They have to choose. Most kids press Keep on their first one. That’s fine. The point is that they pressed it. The studio architecture made them author the change instead of receive it.

Your kid will, at some point in these 30 minutes, ask you: “Should I keep it?” Have one sentence ready. The one that works best is this: “What do you think? It’s your game.” Say it warmly. Mean it. Your kid will look at the card for another beat and then make a choice. Either way, the choice was theirs.

The right answer to “should I keep it?” is “what do you think? It’s your game.” Said warmly. Meant.

Don’t evaluate the AI’s output. “That looks great” teaches the kid that you’re the judge. “What do you think?” teaches them that they are. Carol Dweck’s decades of research on praise comes down to a simple finding: praise the kid’s thinking, not the outcome. “You decided to make the dragon red — that’s a bold call” teaches the kid that decisions are the work. “The dragon looks cool” teaches them that surfaces are the work.²

If the kid presses Review, the studio shows the diff in slightly more detail. They can change a parameter or two before keeping it. If they press Undo, the change vanishes and they can try again. The thing we want them to internalize, in this seventh minute, is that nothing is permanent. The AI proposes. They decide. Undo is always there.

Minute fifteen: when something doesn’t work.

It will happen. Your kid will ask the AI for something specific and the AI will get it wrong. The character will be the wrong color. The scene will have an extra prop the kid didn’t want. The enemy will move the wrong way. The kid will look at the screen and their face will do the thing.

This is the moment that decides whether they come back tomorrow.

Here’s what we’ve learned from cohort intake: the parents who jump in and fix it (or worse, narrate the fix) lose the kid. The parents who say, in a curious voice, “huh, that’s not what you wanted — what do you want to do?” keep them. The first response treats the AI’s mistake as your problem. The second treats it as the kid’s problem to solve, with you next to them.

Most kids’ first instinct is to press Undo and try again with a more specific prompt. That’s the right move and the studio is built to reward it. The second wizard run is almost always closer to what they wanted than the first. The third is usually right. The iteration habit starts here, in the moment when something didn’t work and they decided to try again instead of giving up.

If they get frustrated, you can say one thing that helps: “the AI gets it wrong sometimes. That’s normal. You’re the one who decides what stays.” This is the AI-literacy frame in one sentence. We’ve written a longer version of it at when AI is wrong and what your kid does about it. For these 30 minutes, the one sentence is enough.

Minute twenty-five: “look at what I made.”

Somewhere in the third quarter of the session, your kid will turn the laptop toward you and say some version of “look at this.” This is the moment the studio is engineered around. It’s also the moment most parents accidentally fumble.

The instinct is to compliment the output. Resist it. The kid isn’t showing you the output. They’re showing you the fact that they made something. The thing they want from you is acknowledgment of the making, not evaluation of the made.

What works: ask a real question about it. “What’s the dragon’s name?” “How does the player win?” “What happens if I press this button?” The kid will explain. In explaining, they’ll consolidate the work they just did. They’ll notice things they want to change. They’ll come up with the next wizard they want to run. Your question did the consolidation for free.

What also works: ask if you can play it. If your kid says yes, play it like you would any game. React. Lose on purpose if it’s easy. Find a real moment to be surprised. Mitchel Resnick’s “four Ps” framework for kid-creative environments names Peers as one of the four legs, and in the first session you are the peer.³ You’re the audience that makes the work real.

If your kid finishes the session having shown you one thing and explained it, the session worked. They’ll close the laptop wanting to come back tomorrow, because there’s a thing they’re building and you saw it. That’s the whole game. The license, the AI, the wizards, the dashboard — everything else is in service of that one moment.

If you want to read more about the parent posture across longer arcs, our piece on how to talk to kids about AI without being weird about it covers the conversation side. For the “what you see as a parent” surfaces, the parents page has the dashboard walkthrough. And if your kid is still on the fence about starting, the mini-studios at tellandshow.ai/try let them try a track without installing anything.