A four year old opens the Bobo app and wants to know what is happening on screen. They cannot read yet. Anything that exists only as text does not really exist for this child. Sound is not an extra layer on top of the text here. It is how the story arrives.
Bobo is made for children who cannot read yet
Bobo is aimed at children aged three to five. At that age children listen to stories, they do not read them. They recognise a voice sooner than a word. If you make something for this group, you are making something to listen to.
That shapes everything in the app. A new story without a voice is an empty screen to a four year old. A game whose instructions can only be read is a game nobody understands. Every piece of content needs sound, and it needs it straight away.
The bottleneck: every new line waits for a recording
As long as every sentence has to be recorded in a studio, the studio calendar decides what the app can do. You feel that in several places at once:
- A new story is ready but cannot go live until a recording day has happened.
- Small changes are expensive. Changing one word means booking again.
- Short pieces of text such as buttons, tips and instructions often get no voice at all, because they do not seem worth their own session.
- Reacting to the moment, a holiday or the weather, does not work when you have to plan ten days ahead.
The result is that part of the app has text and no sound. Which is exactly the part the audience cannot use.
One voice, and always the same one
Young children are strict about voices. They hear immediately when something is off. A story read by a different voice than last time does not feel like the same character. At this age recognition is a large part of both the fun and the trust.
So it could not be just any friendly voice. It had to be Bobo's voice, and it had to stay identical across every clip. Not roughly the same. The same.
How the voice clone works
A voice clone was built for Bobo: a model trained on existing recordings of the voice, so that new audio can be produced in that same voice. In the app that works like this:
- A new piece of text comes in, for example a story or an explanation for a game.
- Audio is produced for it in Bobo's voice.
- The clip is listened to and approved before any child hears it.
What used to need a recording day now fits in the short gap between writing something and publishing it. The question shifts from "can we get audio for this" to "what do we want to say".
Careful with a voice, and with young listeners
Cloning a voice is not something you do lightly. Behind a character like Bobo there is a voice actor, and that voice belongs to a person. A voice clone should always come with permission from the voice actor and clear agreements about what the voice may and may not be used for.
Content for young children adds another layer. This group cannot judge what it hears. So a person still listens before a clip reaches the app. Speed is useful, but deciding what a four year old gets to hear stays human work.
The wider lesson: start from what your audience can actually do. Bobo's audience cannot read yet, and at that point sound is not a finishing touch. It is the product.
Further reading
Want to see what Content-Agent makes for your brand?
We load your brand rules and styling once. After that you make images, video, audio and text in minutes, always on brand.