95% of brain development occurs by age 6 - before kids are in school, before they can read, and when screens are discouraged. Effectively, computers aren't accessible during the most important period of child development. Importantly, this is also when a child starts to build social skills. Think of character skills as habits that are formed early in life thru interactive experiences and by modeling behavior. Obviously, it's easier to teach a good habit than change a bad habit so we wanted to be able to not only reach kids early in life but also be able to reach any child, any where. As parents, we know that kids learn to speak before they learn to read and all kids love stuff toys so we made an soft, interactive, and engaging toy that uses Amazon's Alexa platform.
What it does
Say the Magic Words is an engaging, interactive experience to help kids build social skills in a way that was previously impossible. We do this by enabling kids to practice modeling behavior (e.g., when someone says thank you, does the child say, "you're welcome"). Importantly, we focus having the child teach the robot - not the other way around. In general, the child is the one who is modeling the behavior - not the robot - and we do this by giving direct and indirect prompts. One way to gauge proficiency is if a child can teach someone else so we enable that to happen.
How we built it
We built this game several different times, several different ways. We've used Node.JS, Python, and tools. We started out with a design document that listed all the prompts (responses) and user intents. We tested the design of the game with kids and learned A LOT about how to how to engage them while be careful to not over-stimulate (it's easy to stimulate kids...it's actually hard to engage them with educational content). We found a way to make Polly work and optimized the voice for clarity. We also created our own sound effects. Eventually, we built test cases, user flows, etc.
Challenges we ran into
We ran into a lot of challenges using the ASK that had to do primarily with earning how to develop voice-first applications. Early on, we ran into issues w/ natural language of processing of specific words and duplicate intents. Later, we had ran into issues when intended to maximize playabilty (e.g., random X number of scenarios, making changes to sounds, etc.)
Accomplishments that we're proud of
We're proud to have developed an interactive game that kids love and that is actually educational. We're working with a number of schools that comprise more than 100 sites - including a top 10 early childhood provider, large K-12 public district, National School of Character, federal Promise Neighborhood, etc. We also have a community of parents that have been helping us to get better. In doing so, we've received or been nominated for various recognitions (e.g., Harvard Business School's New Venture Competition Regional Finalist, MN Cup Semi-Finalist, MN Coolest Companies, etc.).
What we learned
We've learned that it was hard to incorporate all of our design principles by building using the ASK. We've found that it's better to prototype, test, build, iterate.
What's next for Say the Magic Words
Ultimately, we're looking to move the needle in education by helping kids develop skills that predict success in school and in life. We're piloting with a bunch schools = and exploring a partnership with at least one branch of the government - to prove that not only voice can be an interface for learning but also that we can build and measure social skills in an objective, scalable way.