AAC modelling for parents: what it is, how to do it, and why quizzing kills it

Modelling is what makes an AAC system actually work. Devices don’t teach kids to communicate. Apps don’t teach kids to communicate. People modelling on the device, in real moments, without turning it into a quiz — that teaches kids to communicate. This is the long version of what we coach every AAC family on, written for parents and caregivers who’ve just been handed a shiny new tablet and have no idea what they’re supposed to do with it.

What “modelling” actually means

In AAC, modelling means you use the device alongside speech, in your own communication, in moments your kid is already paying attention. You’re not asking them to use it. You’re not prompting them to use it. You’re showing them what it’s for by using it yourself.

If you spoke only English to a kid for the first three years of life, they’d learn English. If you spoke English but also signed every fifth word, they’d start picking up some signs. If you spoke English and never signed, then handed them a sign-language dictionary at age 5 and asked them to use it, they’d learn nothing. AAC is the same. The system has to be in the language environment before the kid can be expected to use it.

The numbers most parents are never told

Research on aided language input (the clinical term for modelling) suggests kids need to see language modelled on their AAC system for roughly the same amount of exposure they’d need in speech to learn a word. That’s a lot. Some estimates put it at hundreds of exposures per word for genuine, generative use.

For most families, the rate of modelling needed is something like:

• Year 1 of AAC: model at least 1–2 words per minute of interaction, in the moments your kid is engaged. Don’t worry about full sentences. Don’t worry about grammar.
• Year 2+: model in fuller utterances — short phrases, then sentences — at whatever level is slightly ahead of where your kid currently produces.

This sounds exhausting. It is exhausting at first. It gets faster. After about six weeks of deliberate practice most parents do it almost automatically.

The single biggest mistake: quizzing

The most common way AAC implementation goes wrong in homes is quizzing. The parent picks up the device and says: “What’s this? Can you tell me? Press cookie. Press cookie.”

Quizzing has three problems:

1. It’s not how speech is taught. No one walks up to a 14-month-old saying “say milk. Say milk. SAY MILK.” (Or if they do, the kid learns nothing from it.) Kids learn by hearing language used purposefully, not by being interrogated about it.
2. It associates the device with pressure. AAC users who’ve been quizzed for years often become less willing to use their device. The thing that’s meant to be liberating becomes the thing that summons demands.
3. It teaches the wrong concept. Quizzing teaches that the device is for answering questions adults already know the answer to. Communication is the opposite — it’s for sharing things adults don’t know.

The rule we give families: say less, model more, ask almost nothing. If you find yourself about to ask a question with a known answer (“what’s that?”, “what colour is the ball?”, “can you say more?”) — don’t. Just model instead.

What modelling looks like in practice

A morning example. You and your 3-year-old are at the kitchen table. They want yoghurt.

Without modelling: “Use your words! Tell me what you want. Press eat. Come on. Press eat.”

With modelling: You pick up the device yourself and tap “want”, then “yoghurt”, while saying those words aloud. You hand them the yoghurt. You don’t wait for them to do anything. You don’t prompt. You don’t ask them to repeat it.

Tomorrow you do the same thing. And the day after. And the day after.

Somewhere around exposure 200 (or 1000, depending on the kid) your child starts to use the device themselves — sometimes for the same word, sometimes for something completely different that they’ve worked out from watching how the system works.

The principle: model slightly above their current level

This is borrowed from speech development. If your kid is mostly producing single icons, model 2-word combinations. If they’re combining two icons, model three. You’re showing them the next step is possible, not jumping six steps ahead.

You also don’t have to model on every word you say. Realistic targets:

• 1 model per minute is a great start
• Most natural moments to model: requests, comments, refusals, feelings, asking for help
• Most natural not to model: long abstract sentences, internal monologue, things outside the child’s current world

When to use speech, when to use the device

You always use speech. AAC doesn’t replace speech for the people around the AAC user — it adds to it. Adults use both, simultaneously, in roughly the same sentence: speak the word and tap the icon at the same time. This is called simultaneous modelling or aided language stimulation.

For the AAC user themselves, the rule is: whatever they use is communication. If they speak, that’s communication. If they tap an icon, that’s communication. If they pull your hand, point, sign, gesture, or use a single approximation — all communication. AAC isn’t a “replacement for speech”. It’s part of a multimodal system, and the goal isn’t to push the kid toward speech and away from AAC. The goal is robust language access.

What to do when your kid hits the device randomly

Often called “stimming on the device” — and the answer is not to take the device away. Random exploration of icons is how AAC users learn what the system contains. It’s the equivalent of a baby babbling. You’d never tell a babbling baby to “use real words”. You don’t tell an AAC user to “use real icons” either.

Respond to the random hits as if they were intentional. The system says “blue ocean swim” while your kid is in the bath? Great — say “yeah! blue! ocean! swim!” and tap those icons back. Even if your kid didn’t mean it, you’ve just modelled four words in context.

How to handle setbacks

AAC implementation almost always plateaus around month 2–3. The novelty wears off, you get tired, the kid uses it less. This is normal. The fix is rarely “switch to a different app”. The fix is usually:

• Keep modelling. Even when it feels like no one’s watching.
• Move the device into more moments. AAC needs to be in the kitchen, in the car, in the bath, at bedtime — not just at “AAC time”.
• Lower the bar for what counts as communication. Single icons, navigation attempts, even declining to use it — all communicative acts. Celebrate them all.
• Get the rest of the household modelling too. The kid needs at least one fluent adult modeller, ideally two or three. Grandparents, partners, older siblings, the babysitter — all part of the team.

What we coach parents on, session by session

If you work with us on AAC, the first 6–8 sessions usually cover:

1. What modelling is (you’re here)
2. Where your device is — getting comfortable navigating to core words quickly
3. Building a core word vocabulary in your daily routines — picking 10–20 high-leverage words you’ll model in real moments
4. Reading AAC books together — books where the adult models the AAC version while reading aloud
5. Modelling with refusals and protests — “no”, “don’t want”, “stop”, “all done”. These often unlock more genuine use than nouns do.
6. Troubleshooting the plateau — what to do at week 8 when it feels like nothing’s working
7. Bringing other adults onto the team — partner, school, grandparents
8. Long-term language growth — moving from single icons to phrases to sentences without pressure

Final principle

The kids who become fluent AAC users grow up in homes where the adults are fluent AAC users. That’s the work. It’s not glamorous, it’s not fast, and there are no shortcuts. But it’s also not complicated. Model. Don’t quiz. Keep going.

For the funding side of AAC, see our NDIS funding guide. For getting started with a device or app, see starting with AAC. For working with us, book a free 15-minute call.