McCombs Business School, UT Austin
Alexa Stories
Alexa app that teaches children English through stories and conversations.
Overview
Voice based assistants like Amazon Alexa, Google Home, or Siri, are becoming common in our lives, assisting us better and better each day. In this project, our team designs and builds an Alexa app that acts as a conversation partner and can help children of age 5-12 learn the English language. This is an educational app focused on areas with limited educational resources.
Duration
4 months (Jan 2019 - May 2019)
Guide
Prof. Kishore Gawande
My Role
User research, Interaction design, Coding
Research shows that interacting with other native speakers is the fastest way to learn a new language. There are many language learning programs that pair you with a native speaker, or have lessons built on the same concept ex. Pimsleur, Rosetta Stone or Babbel.
Goal
How might we leverage Alexa in helping children learn English in areas with limited teaching resources?
Research
Users
Children aged 7-14
What is a conversation?
Integration with curriculum
With brainstorming around the idea of a stand-alone app, we converged and finalised a roadmap for an app which can be integrated in the school curriculum. This would allow for fewer additional resources, more adoptability, and greater overall impact and learning.
This was based on the idea that instead of just making Alexa a conversation partner, we looked to create an environment of English speakers, where children and learn and grow from each other too along with Alexa.
I created a 3 day teaching plan for the app, which aimed at developing different skills over 3 days.
Designs
Making conversations human
User testing
We conducted usability tests with our target audience, children, in a classroom setting. Key takeaways were:
Response time of children was less than planned due to excitement. So spoke in between. Better planning for dialogue and error preventions needs to be done.
Alexa sometimes got confused due to multiple voices. Error prevention needs to be improved.
Children lose interest if Alexa goes into "Sorry, I cant understand" mode for long.
Story based quiz approach was loved by children.
Children came up with more variety of correct answers than had been accounted for in test run.
With brainstorming around the idea of a stand-alone app, we converged and finalised a roadmap for an app which can be integrated in the school curriculum. This would allow for fewer additional resources, more adoptability, and greater overall impact and learning.
This was based on the idea that instead of just making Alexa a conversation partner, we looked to create an environment of English speakers, where children and learn and grow from each other too along with Alexa.
I created a 3 day teaching plan for the app, which aimed at developing different skills over 3 days.
Reach out for full case study
Bring your own data
Easily integrate and structure your own data to be used in research through no-code visual tools.
Reach out for full case study
Package, deploy, and monitor ML models
No-code, click-through experience for packaging and deploying models without needing IT, software engineering, or DevOps knowledge.
Reach out for full case study
User Research
Personas
The platform is based on standardization of three core components of a machine learning model: Data, Code, and Pipelines.
User needs
The platform is based on standardization of three core components of a machine learning model: Data, Code, and Pipelines.
