When a family decides to move to a communication app, they run into dozens of options, prices ranging from nothing to several hundred euros, and a great deal of technical jargon. The question I get asked over and over is always the same: “which one is best?”. And the honest answer is that there is no best app — there is the one that fits this particular person, in their particular setting.

This guide gathers the criteria that genuinely make a difference day to day, ordered by how much they weigh in practice.

Before looking at apps: three questions about the person

Choosing a tool before understanding the need is the most expensive mistake. These three questions save months of trial and error.

How are they going to select? A child who points precisely with their index finger is not the same as one with motor difficulties who needs large, well-spaced cells, or one who drags a finger across the screen. Cell size and tolerance of accidental touches matter more than the size of the pictogram library.

How much vocabulary do they handle today? An app showing hundreds of symbols at once overwhelms someone who is starting out with five basic requests. And the other way round: an app with six-cell grids runs out of room within months for someone already combining words. You need something that works today and in a year’s time.

Where are they going to use it? At home, in the classroom, at the park, in the car, at a medical appointment. The answer shapes every other decision, especially the one about internet access.

The criteria that do matter

1. It has to work offline

For me this is the decisive criterion, and the one most often overlooked when comparing. A communication app that needs the internet in order to speak is an app that falls silent exactly where it is needed most: in the playground, on a school trip, in a waiting room with no signal, in a building with overloaded wi-fi.

Communication is not an optional service that can wait for the signal to come back. Check explicitly that the pictograms and the voice work in aeroplane mode, not just that the store description says it “can be used offline”.

2. The voice

Listen to the voice before deciding, and listen to it with a long sentence, not a single word. Many apps use the robotic voice that ships with the operating system, and there is an enormous difference between that and a natural voice.

This is not an aesthetic question. A voice that sounds artificial makes the people around — classmates, the waiter, a grandparent — take longer to understand the message, and sometimes they do not even realise they are being spoken to. The voice is the part the rest of the world hears, and it determines whether communication works socially or stays an exercise between the person and their tablet.

3. Real customisation

Every app is sold as “customisable”. What you need to look at is how far that goes:

Can you add your own photos? The actual cup at home, grandma’s face, the family dog. Generic pictograms are not always enough.
Can you record your own voice, or a family member’s?
Can you rearrange the grid so the most-used words are always in the same place? Positional stability is key: the person communicating memorises where each symbol is, and an app that reorders cells on its own destroys that learning.
Can you hide what is not needed yet without deleting it, so you can switch things on gradually?

4. How many profiles, and who manages them

If you are a professional working with several students, a single-profile app forces you to reconfigure everything between sessions. If you are a family, you may want the school to be able to see or replicate the setup used at home, so the system is the same in both settings. Consistency between home and school is one of the factors that most accelerates learning.

5. Privacy

This gets checked rarely and should be checked always. A communication app records the most intimate thing there is: what a person wants to say. Before installing, look in the privacy policy for whether the history is kept on the device or uploaded to servers, and whether it is shared with third parties.

My rule is simple: communication data should not leave the device. If an app is not clear about that, it is reason enough to rule it out.

6. The price and, above all, the pricing model

It is worth telling three models apart, because their long-term impact is very different:

Free: perfect for trying things out at no risk. Check whether they are genuinely free or whether the important features are locked.
One-off payment: you pay once and it is yours. Usually works out better for years of use.
Subscription: a monthly or annual fee. Here you need to do the long arithmetic, because a communication system is not used for three months, it is used for life. And ask yourself an uncomfortable question before subscribing: if you stop paying one day, is the vocabulary built up over years lost?

Common mistakes when choosing

Choosing by number of pictograms. An app with 10,000 symbols is not better than one with 3,000. Nobody uses 10,000 symbols; what matters is finding the ones you need quickly.
Deciding without the person who will use it. Try it with them before buying. Their real reaction to the first two sessions is worth more than any review.
Switching apps every few months. Every change means relearning where everything is. An acceptable app used consistently beats the perfect app started afresh each term.
Buying it and not training the people around. The app does not communicate on its own. If the family and the professionals do not use it by modelling — pointing at symbols while they talk — it ends up in a drawer. This is by far the most common reason people give up.

A quick list for comparing

Once you have two or three candidates, answer this for each of them:

Does it speak without internet, in aeroplane mode?
Does the voice sound natural in a full sentence?
Can I add my own photos and my own words?
Do the cells stay where I put them?
Does the history stay on the device?
What happens if I stop paying?
Can it grow with me from 5 symbols to full sentences?

If an app answers these seven well, it is a serious candidate, whatever it costs. If it fails the first or the fifth, I would rule it out however good the rest looks.

Where ComunicaTEA fits

I built ComunicaTEA from exactly these criteria, after seeing in schools what was failing with the existing tools: communicating works offline, it has around 1,000 words with Mulberry pictograms, it says sentences out loud, and its prediction suggests the next word based on what each person uses, without ever moving the board. Prediction is calculated on the device, and the history and statistics are stored there; only what you turn on leaves it, such as natural voices.

It is not the only valid option, and I do not claim it is. But if your list of requirements looks like the one in this article, it is worth a look.

Conclusion

Choosing an AAC app is not about finding the most complete one, but the one that person will be able to use every day, everywhere, for years. Start by understanding how they select, how much vocabulary they handle and where they will use it; insist that it works offline, with a natural voice and with the data kept safe; and above all, support the use of it. The tool matters, but what really lets someone find their voice is the people around them who respond.