Alexa Kids Skills: Teaching Children to Read with Alexa

We’re no strangers to Alexa Skill development, having built a number of Skills on behalf of our clients and in our team’s own time. The one thing we’d never done before? Use Alexa in the field of education.

Having previously collaborated with one the best storytellers around; Wonderbly, who design beautifully unique children’s storybooks, we knew the impact voice assistants can have in engaging with children on a more interactive platform.

Especially when it comes to what’s been coined as “the Alexa generation“.

Every week a member of our team is working on an innovative project or prototype to solve a problem they’re passionate about.

This week we challenged designer James and developer Kev to explore the role of voice assistants in teaching children how to read.

With experience in conversational design and Alexa Skill development, we knew they had the experience to bring the project to life. Here’s how we did it.

With only five days to solve the challenge, we knew that developing an entire platform was out of the question. In order to prove the concept of an Alexa Skill teaching children how to read, we had to achieve a few things by Friday:

• Create a user journey that allows children to progress and develop their reading skills
• Design voice interactions easy enough for children to understand and use
• Develop a feature that teaches children to read and write different words.

Knowing that James had previously designed conversations, and Kev had developed voice interactions, our biggest challenge was obvious; we’d never actually taught anyone how to read. Thanks to a little input from the team, we knew that using phonics was the best way to go.

Phonics are a method of teaching children how to read by correlating sounds with symbols; breaking down words into the individual sounds they contain, and letting children recognise these patterns and combination of sounds to read full words and phrases.

The Alexa Skill needed to use phonics in a feature, or series of voice interactions, so that at the end of every page in the story, Alexa could prompt a ‘repeat after me’ line of dialogue, getting children to sound out words correctly before reading them in full and progressing to the next page in the story.

After opening the Skill and starting the story, Alexa welcomes users and checks if they’re ready to get started. There’s also the option to open the Skill on a certain page or continue from where users left off the last time they were reading.

There are a few different responses from users that prompt a different reaction from Alexa; reading words out correctly and moving onto the next page, getting it completely wrong and having a re-cue for a second attempt, and getting close to the word but mispronouncing certain sounds within it.

If a user does get close to the word, Alexa will break it down into phonics and use a ‘repeat after me’ method that can teach children individual sounds, before putting them together to complete the word or phrase.

These interactions on pages get more and more difficult as children develop their skills and learn sounds that can be used across different words. Using lines of dialogue throughout that encourage children struggling and reward them for doing well.

Having spent the week trying to solve the challenge, James and Kev presented Story Circle back to the team, who discussed their ideas and experience, gave feedback to the guys and came up with a few suggestions for the future of the project:

Phonics are difficult to get your head around, in fact, many schools actually hold training sessions for parents so they know how to use them correctly at home while teaching their children how to read. To assist the voice-led platform, an app could be used by parents and children to tackle more difficult phonics, using this visual way of learning to support the more engaging and entertaining story on the Alexa Skill.

One thing that stood out was the default Alexa voice, and even the Amazon Polly alternatives, feeling robotic and almost cold to interact with. We do, however, have the ability to record lines of dialogue and programme these as the responses in the Alexa Skill, which is something we definitely plan on exploring in the future (it’s very time/resource intensive).

We also had the idea of measuring different areas of learning that children particularly struggle with. By identifying key sounds, words and phrases that every individual child is having a tough time learning, pages of the books and interactions could be personalised to help them overcome the exact words or sounds they’re stuck on.

To aid our team in building a simple story for Alexa, Kev developed a CMS to quickly write up pages in a story, design user flows and develop core voice interactions.

Using the CMS to translate stories into SSML:

SSML (Speech Synthesis Markup Language), is the language used by developers that’s understood by Alexa Skills. With an easy translation system for every page in our story, we could tell Alexa what to say, when to pause, which words to use when interacting with users, and which phonics and sounds each of these words contains.

Turns out we’re a pretty young team (28 years on average). 

But thanks to the parents in the room, we immediately knew the role phonics play in teaching children how to read. Breaking words down into different sounds to make them easier to say, read and write. The potential phonics offer in more interactive and engaging learning methods (such as songs and games), translated beautifully into an Alexa Skill that uses the same methods to teach children how to read.

Exploring the different phonic teaching methods:

We explored a few different teaching methods that have been used in schools to help children learn to read, which mainly come down to three areas; Analytical phonics, which are used to match the sounds in a known word with the sounds in an unknown word, linguistic phonics, which focus on using repetition with similar sounding words like cat, hat and bat, and synthetic phonics, which blend individual sounds together to create the full word or phrase children are trying to read.

If children get close to reading a word or phrase correctly, Alexa will prompt them with similar sounding words before they give it another go. The ‘repeat after me’ method means that children can be taught more difficult words and phrases by the individual sounds within the word, before progressing through the Skill and the story.

We didn’t want to contradict teachers or schools:

One crucial element of the project was making sure that the Alexa Skill uses the same phonic teaching methods used in schools, ensuring that educating children outside of normal lessons didn’t teach them bad practices, or ways of learning that would contradict what teachers are also telling them.

With a working prototype in place and a plan to develop the project in the future, we’ll spend the next few weeks of Innovative Time finessing the Alexa Skill and developing the additional features and functionality highlighted by the team.

We’re still not experts in teaching children how to read.

What we do know is that to bring the best projects to life, we need to collaborate with organisations that are experts in their fields. If you have experience in children’s education, we’d love to collaborate.