Our team has been creating and publishing word games for more than 10 years now. Some of our super-hit games like Lexulous have been covered in leading publications like the Wall Street Journal, New York Times, New Yorker and Financial Times. As on date we have over 8 million players of our games.
While developing these games we have traditionally focused on the visual user experience.
In recent years there has been rapid development in voice recognition and natural language processing technologies. Our experiments with products like Amazon Alexa have shown that voice interactions are natural and easy. The technology is also conducive for third party developers to build on.
Voice Recognition is a huge opportunity for developers. It opens up the untapped market of providing services & entertainment to our visually impaired friends (285 million people across the world as per WHO!). Developers can use this technology to bridge the gap between the visually impaired and those who are not.
So, our inspiration is to take this technology forward and:
- Create meaningful entertainment experiences for the visually impaired.
- Allow the visually impaired to connect easily with their friends & family.
What it does
This particular skill allows the user to start a game of anagrams. In each round the user needs to solve 3 anagrams. Points are awarded based on how easy or difficult the anagram was. The game keeps track of the user's score over a period of time and the user is informed of their global rank whenever they start playing the game.
The game has a useful feature that allows the user to ask for a hint in case they are stuck somewhere. The game shall define the anagram for the user and allow them to improve their vocabulary too!
It also contains a feature that you can use to challenge a friend! This is really cool and adds an element of social networking to voice interactions. We want to develop this further as the API opens up for sending out messages.
How we built it
Building this game was pretty intensive!
The Alexa Skills SDK is easy to use and our team has extensive experience in AWS Lambda. So working with the technology was not an issue at all.
We started out with a game-play/brainstorming session where each team member closed their eyes and one person was acting as the "game". This session was duly recorded. Afterwards our team members listened to the recording and analysed the voice interactions. This analysis gave us an insight into how we speak and interact with our eyes closed and no devices to touch.
Converting the analysis into key phrases and making the interaction with Alexa was the time consuming part of our work.
The other challenging portion was to create a system of approximately matching user utterances to the actual values we were looking for. This was to cater for diverse accents.
Accomplishments that we are proud of
We are proud of having built our first voice based game. This experience has been amazing and has taught us to think in a completely different manner. Traditional programming works by giving the user a limited set of functions to react to. When we start using Voice, then the user is open to saying anything they want. This makes it all the more important to properly guide the user and create a meaningful experience, all the while providing the user full flexibility of interaction.
What we learned
Our team is now confident of developing voice based interactions. We understand how the technology works, how Alexa Skills work and what can be achieved out of this. We also learnt that it is important for each voice based project to have its own personality. Using the SSML tags available in Alexa it is possible to greatly expand on it and have a unique touch to your work.
What's next for Voice Based Word Games
We are looking to expand the social networking aspect by adding more features to the user challenge module. Among other features, this includes sending messages and head to head stats.
Our team is now going to add more games to it. To do this we would need to create a structured menu that can be easily navigated via voice commands. This is a bit challenging as we need to ensure that the user is aware as to which sub-level menu they are in.
We are also actively looking at using more of SSML to create a unique personality for our products. So that whenever you use any of our Voice Products you will automatically understand that its from our team!