Build a Recommendation Engine With Sticky Notes (No-Code AI Demo for Ages 7–12)

What Kids Are Really Learning (and Why It Matters)

If your child has ever said, “How does YouTube know what I want to watch?” or “Why does this app keep suggesting the same kind of games?”—they’re noticing something big: recommendation systems.

In grown-up terms, recommendation engines are a type of AI that suggest what you might like next based on patterns. In kid terms: a smart helper that guesses your favorites by paying attention to what you pick.

This sticky-note activity is an ai activities without coding demo that makes those ideas visible. Kids don’t need a computer, and you don’t need to “teach AI.” You just guide them through a few rounds of choosing, sorting, and predicting.

By the end, your child will have practiced key machine learning concepts for kids, like:

• Data: the choices we record
• Features: the “ingredients” that describe an item (funny, animals, adventure)
• Patterns: what shows up again and again
• Prediction: making a best guess about what someone will like next
• Feedback loop: the system improves when it sees more choices

And the best part: it feels like a game.

What You’ll Need + Setup (10 Minutes)

This is a simple stem activity recommendation engine you can run at the kitchen table.

Materials

• 20–30 sticky notes (3 colors if possible)
• 1 marker
• A flat surface (table or wall)
• Optional: stickers for “likes”

Step 1: Choose your “items” to recommend

Pick a category your child actually cares about:

• Books you own
• Movies you’ve watched
• Snacks in your pantry
• Weekend activities (park, crafts, bike ride)
• Games or toys

Write one item per sticky note. Keep the names short: “Animal Book,” “Space Movie,” “Pretzels,” “Art Time,” etc.

Step 2: Add 3–5 “feature tags” for each item

A feature is a describing word that helps the system understand why someone might like it.

Good feature tags for ages 7–12:

• Funny
• Adventure
• Animals
• Science
• Sports
• Calm
• Creative
• Strategy
• Sweet
• Crunchy

Write 2–3 tags in small text on each sticky note (or put tiny mini-sticky notes on top).

Step 3: Make a “Profile Area” for each person

Make one space on the table per family member (or at least “Kid” and “Parent”). This is where their “liked” items will go.

The Sticky-Note Recommendation Engine (3 Rounds)

This is the heart of recommendation systems explained for kids. You’re going to run a mini version of what real platforms do: collect signals, find patterns, and recommend.

Round 1: Collect data (Likes)

1. Spread all item sticky notes face-up.
2. Each person picks 3 items they like and places them in their profile area.
3. For each liked item, copy its feature tags onto a “Feature Score” space.

How to “score” without math stress:

• Each time a feature appears in something they liked, add a tally mark next to that feature.

Example: If the child likes “Animal Book (Animals, Funny)” and “Zoo Trip (Animals, Adventure),” then Animals gets 2 tallies, Funny gets 1, Adventure gets 1.

Round 2: Make recommendations (Predict)

Now you (or the child) become the recommendation engine.

1. Look at the top 2–3 features in the person’s profile.
2. Find 2 new items they haven’t picked yet that match those features.
3. Present them as recommendations: “Based on your picks, I recommend…”

Then the person responds:

• Like: add it to their profile area
• Skip: move it aside

This is the feedback loop.

Round 3: Improve the engine (Update)

Repeat one more time:

• Add feature tallies from the new “liked” recommendations
• Make 2 more recommendations

Kids quickly notice: the more data you collect, the better the guesses get.

A simple scoring method you can use (kid-friendly)

If your child likes numbers, you can introduce points—still no heavy math.

• Each matching feature = +1 point
• Recommend the items with the most points

This turns the activity into a mini puzzle.

Example: A Ready-to-Use Mini Dataset (Copy This)

If you want to start fast, use this sample “Weekend Activities” set. Write each row on a sticky note and include the features.

Item (Sticky Note)	Feature Tags (2–3)	Best For Kids Who Like…	Parent Tip
Bike Ride	Sports, Adventure	movement, outdoors	Add “short/long” as a feature if energy levels vary
Baking Cookies	Creative, Sweet	making, tasting	Feature idea: “Messy” vs “Neat”
Science Experiment	Science, Creative	curiosity, building	Keep a “safe materials” rule
Board Game	Strategy, Calm	thinking, teamwork	Add “Competitive” as a feature
Nature Walk	Calm, Adventure	exploring, collecting	Bring a bag for “treasures”
Drawing Time	Creative, Calm	art, quiet focus	Offer prompts: animals, space, comics
Soccer Practice	Sports, Social	groups, running	Add “Indoor/Outdoor” as a feature
Library Visit	Calm, Science	books, learning	Let the child be the “recommender” for you

Want to customize? Let your child invent features. Some of the best ones are surprisingly honest: “Silly,” “Too Loud,” “Cozy,” “Hard,” “Fast.”

Talk Like an AI Mentor: Questions That Make the Concepts Stick

As you play, sprinkle in a few questions. This is where the machine learning concepts for kids land naturally.

Try these:

• “What clues did the engine use to guess your next favorite?”
• “If we only knew one thing about you, what should it be?” (teaches strongest features)
• “What happens if you pick something random just once—does it confuse the engine?” (teaches noisy data)
• “Is it fair if the engine only recommends one kind of thing?” (teaches filter bubbles)
• “What feature should we add so recommendations feel more accurate?” (teaches better data)

A quick, kid-friendly explanation of two common recommendation styles

You can demonstrate both using sticky notes:

• “People like you also liked…” (collaborative filtering)
  • Compare two profiles: if you and your sibling both liked two of the same items, try recommending your sibling’s other favorite.
• “You like X, so you might like Y…” (content-based)
  • Recommend based on shared features: you liked “Animals” twice, so here are more “Animals” items.

No need to name them during the activity—but if your child is curious, these labels are a nice “bonus level.”

Make it more challenging (Ages 10–12)

Add one of these twists:

• Cold start problem: Try recommending after only 1 like. Ask: “Why is it harder?”
• New item problem: Add a brand-new sticky note with no tags. Ask your child how to tag it.
• Diversity rule: Force one recommendation to include a feature the child hasn’t picked yet (“Try something new”). Discuss how apps balance favorites with discovery.

Next Steps: Turn This Into an Ongoing Family AI Habit

Here’s how to keep the learning going (without turning it into homework):

• Play again next week with a new category: snacks, books, chores, after-school activities.
• Let your child be the “algorithm”: they’ll practice logic, empathy, and pattern-finding.
• Do a “mystery feature” round: hide the tags and let the engine guess them after the like/skip feedback.
• Connect it to real life: next time an app recommends something, ask: “What sticky-note features do you think it used?”

If your child enjoyed this, they’re ready for more structured AI learning that still feels playful. At Intellect Council, we build interactive lessons that take the same ideas—patterns, features, and predictions—and help kids level up step-by-step.

Your simple goal for the week: run this sticky-note recommendation engine twice, and let your child improve it once. That’s real AI thinking, no coding required.